a2p to ott, chat bots to smart future of predictive · introduction mobile messaging now: stats,...

38
SUPPORTED BY CAN A BOT TEACH YOU A LANGUAGE? WHAT HAPPENS WHEN MACHINES TEXT ‘HELP’? HOW A SIMPLE TEXT MAKES SHOPPING SAFER CAN YOU CHAT WITH THE NEWS? PREDICTING THE FUTURE OF PREDICTIVE TEXT powerful communication medium is powerful communication medium is machines … how the world’s most A2P to OTT, chat bots to smart powerful communication medium is machines … how the world’s most A2P to OTT, chat bots to smart powerful communication medium is machines … how the world’s most A2P to OTT, chat bots to smart O H S T O H R E F A S G N I P P O S E K A M T X E T E L P M I S A W O T E V I T C I D E R P F O E R U FUT T G N I T C I D E R P powerful communication medium is W T X E F E H T shaping up for tomorrow ? P L E H T X E T S E N I H C A M N E H W S N E P P A H T A AT H W shaping up for tomorrow H C A E T N A C CA S shaping up for tomorrow A U O Y H T O B A N O H S R E F A S G N I P P O H T H C C CA ? S W E N E H T I W T A AT H U O Y N A G N A L H C A E T ? E G A U G A U O Y H SUPPOR Y TED B SUPPOR

Upload: others

Post on 22-May-2020

4 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

SUPPORTED BY

CAN A BOT TEACH YOU A LANGUAGE?

WHAT HAPPENS WHEN MACHINES

TEXT ‘HELP’?HOW A SIMPLE

TEXT MAKES SHOPPING SAFER

CAN YOU CHAT WITH THE NEWS?

PREDICTING THE FUTURE OF

PREDICTIVE TEXT

powerful communication medium is

powerful communication medium ismachines … how the world’s most

A2P to OTT, chat bots to smart

powerful communication medium ismachines … how the world’s most

A2P to OTT, chat bots to smart

powerful communication medium ismachines … how the world’s most

A2P to OTT, chat bots to smart

OHSTOH

REFASGNIPPOSEKAMTXET

ELPMISAWO

TEVITCIDERP

FOERUFUTTGNITCIDERP

powerful communication medium is

W

TXEF

EHT

shaping up for tomorrow

?’PLEH‘TXETSENIHCAMNEHW

SNEPPAHTAATHW

shaping up for tomorrow

HCAETNACCA

S

shaping up for tomorrow

AUOYHTOBAN

OHS

REFASGNIPPO

HTHCCCA

?SWENEHTIWTAATH

UOYNA

GNALHCAET

?EGAUGAUOYH

SUPPOR

YTED BSUPPOR

Page 2: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE CONTENTS

03

3. MACHINE MESSAGING

A new SIM for machines that text: CLX..........................................................40

Post-human messaging: SAP Mobile Services .......................................42

Pegg is your bot accountant: Sage...................................................................44

Bots for learning a language: Duolingo...........................................................46

Ask a bot about your spending: Mastercard ...........................................48

Can you chat with the news? Quartz..............................................................50

Bots to help people work smarter: Slack .....................................................52

4. WHAT’S NEXT

Is messaging the new runtime? Andreessen Horowitz.....................56

Network insights will make messaging smarter: Telefonica............58

In-house OTT messaging for every app: Layer .......................................60

Conversational commerce – the future? Operator............................62

Payments and messaging apps: Consult Hyperion..............................64

Google, the operators and SMS v2.0:.............................................................66

AI and predictive messaging: Swiftkey...........................................................68

1. INTRODUCTION

Mobile messaging now: stats, trends, challenges ..................................08

The glittering future of A2P: OpenMarket.....................................................10

Enterprise messaging: SAP Mobile Services...............................................12

SMS – still best for 2-factor authentication: CLX....................................14

2. USE CASES

Text-based crowd control: Virgin Trains .........................................................18

How real time SMS alerts are tackling card fraud: FICO..................20

Machines that send ‘help me’ messages: Philips Healthcare.......22

Texting for better schoolwork: Behavioural Insights Team ..............24

Getting smarter with ‘text to donate’: Cancer Research UK........26

Text is driving the drivers: Uber............................................................................28

Rich messaging for feature phones: iTouch................................................30

How chat made payments fun: Venmo.........................................................32

Towards secure and private chat apps: Wire...........................................34

OTT messaging in South Africa: WeChat......................................................36

MEF FUTURE OF MESSAGING GUIDE CONTENTS

02

MEF FUTURE OF MESSAGING GUIDE CONTENTS

03

Page 3: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE FOREWORD

People say that the digital age is destroying ourability to communicate. We’re staring at screens,avoiding conversation. But what are we doing

when we’re on our phones? Well, mostly conversing.

The truth is, people love to communicate. And mobilemessaging is how billions choose to do it. We’ve seenthis in the rise of SMS, which grew ten-fold from 2004to 2014. And now there is a similar explosion in theuse of chat apps like WhatsApp and WeChat.

But it’s not just people who are chatting and texting.Businesses also choose mobile messaging. With goodreason. According to Mobile Marketing Watch, textmessages have a 98 per cent open rate (email is 20 percent). And most are opened inside five seconds. As aresult, the A2P (application to person) market isbooming across all sectors – as enterprises embrace

this trusted, reliable and cost-effective channel. Now,an exciting third trend is emerging: messaging bymachine. Here, smart devices send alerts to signaltheir status, saving firms money and time. Then thereare chat bots - virtual machines that could be thefuture of customer engagement.

So in this guide, we reveal the power of messaging inall its forms. Inside, you’ll learn how companies usemessaging to solve real business problems. We lookat emerging platforms, and explore what’s next forthe medium.

The guide been commissioned by the global tradebody Mobile Ecosystem Forum. It is supported bymembers of MEF’s Future of Messaging Programme,which promotes best practice and innovation acrossthe messaging ecosystem.

Welcome to thefuture of messaging

“Text messages have a 98 per cent open rate. Andmost are opened inside five seconds. This is why theA2P market is booming.”

05

Page 4: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

Person to person messaging via SMS haspeaked, but ‘OTT’ chat apps are attractingbillions of users.

Meanwhile businesses are discovering thepower of text as the world’s most direct formof communication.

In this section, we look at the current state ofmessaging and analyse its key challenges andopportunities.

INTRODUCTION01

06 07

Page 5: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

The mobile phone wasinvented for voice. Butthe people saw itdifferently. Texttriumphed, and nowthe medium is readyfor its next phase.

Personal, universal...The unstoppable rise of text

MEF FUTURE OF MESSAGING GUIDE INTRODUCTION MEF FUTURE OF MESSAGING GUIDE INTRODUCTION

In 2012, Austin Wierschke successfully defendedhis US National Texting Championship crown. The17 year old texted 149 characters in 39 seconds

with no spelling or punctuation errors. Impressive.But how long would it take to speak 149 characters?Actually, about five seconds.

In many obvious ways, voice is better than text. It’scertainly faster. And yet 20 years since people sentSMS messages from user-unfriendly numerickeypads, text is clearly the dominant force in mobile.

Look at the numbers. According to Telefónica, mobileusers currently send around 8.3 trillion textmessages a year. And then there are the businesses.Companies, utilities, broadcasters, hospitals - they’veall embraced the text alert as an affordable andreliable channel for customer engagement. Now, theso-called application-to-person market (A2P) is setto grow from 1.6 trillion text messages in 2015 to1.76 trillion in 2018 (Portio Research).

To repeat, text has triumphed. According to a 2015study by Infomate, Americans spend 26 minutes a daytexting against six minutes on voice calls. So why is this?Part of it may be social reserve. But there are practicaladvantages too. The asynchronous nature of text maybe the most significant. A text does not demand to beanswered immediately like a call. It sits thereunobtrusively, waiting for a response.

People like that. And so does business. Indeed, thebenefits of text to firms are even more compelling. It’subiquitous: anyone with a phone can use SMS – nomatter what handset, operator, OS or country. It’s(almost) always on because messages are transmittedover cellular networks, rather than 3/4G. And peoplerespond very quickly to SMS. According to a Dynmarkreport, SMS has a 98 per cent open rate - and nine inten texts are read within three seconds.

Of course, for all the dominance of mobile messaging,the medium is currently undergoing big changes. The

SMS channel is over 20 years old, and is beginning tobe eclipsed – in the person-to-person space atleast – by the so-called ‘OTT’ apps, which let userssend messages over IP. The apps – WhatsApp,Facebook Messenger, WeChat, Viber and others –give people the ability to send video and photos, andeasily set up closed groups. These are features thatare not available via SMS.

Over to over the topA 2016 study by the Mobile Ecosystem Forumrevealed OTT apps are now favoured by a majority ofend users. For example, 56 per cent of people saythey regularly use Facebook Messenger, against 42per cent for SMS. The shift from SMS to OTT in someuse cases is real, though it should be stated thatperson to person SMS is not in decline, just changing.And in the A2P space the OTT apps have yet to makea serious impact. To re-state the above: SMS isuniversal and ubiquitous; OTT apps are not. Thisclearly matters to business.

Companies derive huge value from the simple abilityto contact customers in a timely fashion.

This guide reveals some apposite examples. ConsiderUber. It’s hugely complex business that couldn’tfunction without real time texts between riders anddrivers. Or FICO, helping banks reduce fraud by textingcustomers at the moment of purchase to discoverwhether a flagged transaction is genuine.

More to comeBusinesses across all verticals already leverage A2Pmessaging. But there’s immense potential for morecompanies to explore the channel. The MEF studyrevealed that almost one in four respondents has yetto receive an SMS from any of the verticals listed.

Meanwhile, new possibilities are emerging such asmessaging inside machines. Connected things needa way to talk to each other and cellular networks offera better route than wi-fi for example. Operators’ newnetworks will consume less power and potentiallyconnect billions of devices simultaneously. They willenable machines to send diagnostic messages foryears without battery issues. And dynamic SIMs willlet them connect wherever they are shipped.

Chatbot comethThe other big change coming to messaging concernsa different kind of machine – a virtual one. The hugepopularity of messaging has led many to ask thefollowing question: if people like texting so much, whynot make it the option for ‘talking’ to digital services.

For example, say you want to find out a train time. Youcould open an app, click on a menu, find a drop downlist of destinations, click on a departure time etc. Oryou could go to Facebook Messenger and just ask anartificially intelligent bot. And it would answer quicklybecause it would know so much about you anyway.

This is happening now. Facebook has opened its botAPI and already has many thousands of apps on itsplatform. The potential for so-called conversationalcommerce is immense. But it’s not the onlypossibility. In this guide, we show how Quartz hascreated a bot that ‘chats’ the news, Duolingo usesbots to help people learn languages and Sagedeveloped a bot accountant. The future of messagingis already here.

* The A2P messaging market will be worth$58.75 billion in 2020 (Mobilesquared)

* Banks, retailers and OTT providers willaccount for 60 per cent of A2P traffic by2017 (Telefonica Text Economy Report)

* 44 per cent of people prefer product andmarketing messages via SMS (DirectMarketing Association)

* Businesses will send over 1.7 trillion A2Pmessages in 2016 (Portio Research)

*Developers built 30,000 bots in six monthson Facebook Messenger

* People sent 8.3 trillion texts in 2016

* By 2018, the number of chat app usersworldwide will reach two billion (eMarketer)

08 09

Page 6: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE TEXT IN THE ENTERPRISE MEF FUTURE OF MESSAGING GUIDE TEXT IN THE ENTERPRISE

In the consumer space, SMS has peaked. So is that the end of the boom yearsfor text? Not a bit of it, says OpenMarket’s GM Jay Emmet. He sees a stellarfuture for SMS in the enterprise.

In the tech world, one thing never changes and it’sthis: everything changes. The industry, its fans andits media are always looking for the next big thing.

Fair enough. New things are exciting. The risk is tooverlook the value of what we already have.

And so we come to SMS. The veteran mobile mediumis now over 20 years old, and is often unfavourablycompared with the new OTT messaging apps. Thelatter services can exchange rich media, set up closedgroups, support gifs and a lot more. What can SMSdo? Not as much. But that’s the whole point. Textscan ‘only’ be read – but anywhere, and by anyone.

Jay Emmet, GM of messaging specialist OpenMarket,loves these limitations. He’s worked with SMS fortwo decades and his enthusiasm is undimmed. “Factsare facts. Yes, its true SMS was born in the 90s, butits engine is proven, it’s very useful and it’s not goingaway,’ he says. “The point about SMS is that it’s inevery handset and every network on the planet. Andevery human being is a trained user – there’s noeducation process needed. SMS has a ubiquity that’sunique. If you have a global or even regional need foromnipresent reach, there is no substitute for SMS.”

Testing, testingOf course, plenty of companies have a ‘need foromnipresent reach’, which is why Emmet is so bullishabout the future of SMS in the enterprise. That said,the so-called A2P (application to person) market fortext has been around for years. Why is it taking solong? Emmet believes it comes down to corporateinertia. He says: “Big corporations are conservative.You suggest to a Fortune 500 that they switch fromemail to text, and it’s provocative. It takes them a longtime to be convinced. Also, when they do commit,they’ll test it first. (But) I really think SMS will becomethe preferred enterprise channel. And I predict that in

three or four years more texts will be sent bybusinesses than consumers.”

While Emmet argues for the ubiquity and universalityof text, he also believes the medium has another keybenefit: it is asynchronous. “Compare the experienceof waiting on a voice call for 15 minutes with waiting15 minutes for a text reply. People don’t mind waitingwhen it’s under their control. That’s not the case witha voice call. And while my generation will put up withthose calls, the new generation just won’t,” he says.

Engineering ROINeedless to say, when customer care arguments aremarried to demonstrable cost savings, that’s whenenterprises sit up and listen. Emmet gives theexample of one client, a cable company with a fleetof installation trucks. Every time its engineers missan appointment because the customer is out, it costs$200. So when the company started to send textshalf an hour before asking if people are home, missedcalls fell by ten per cent.

Emmet says these savings are huge. “We’re talkingabout a company that works hard in other areas ofits operation to cut costs by a fraction of a per cent,”he says. “They haven’t just cut costs, they’ve alsoimproved customer experience. Normally, you do oneof those at the expense of the other.”

This example reveals another reason why text can beso compelling: people always read them – typicallywithin a minute. It’s the best channel for anythingtime sensitive, though some companies have beenslow to realise it. “We have an e-commerce customerthat sends emails when transactions fail,” saysEmmet. “They send these emails – and no one opensthem. So now they have started to send texts toremind people to look at their emails!”

Enterprises on alert

10 11

“ In three or fouryears more texts

will be sent bybusinesses than

by consumers.”

Jay Emmet, GM of OpenMarket

Page 7: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE ENTERPRISING IDEAS MEF FUTURE OF MESSAGING GUIDE ENTERPRISING IDEAS

Adding a mobile messaging function to enterprise systems is a must forbusinesses, says Mike Willock, VP for EMEA Sales, SAP Mobile Services...

Integrateexpectations

For any enterprise embracing digital to reachcustomers, mobile and messaging is a keycomponent and enabler. It’s why enterprise

software giant SAP bought Sybase 365 (now SAPMobile Services). The acquisition opened up new,broader messaging and integration opportunities withexisting products and as standalone solutions.

Mike Willock, the division’s VP for EMEA Sales,explains: “Messaging continues to be one of the bestchannels for reach, simplicity, reliability and security.It offers enterprises an inexpensive yet highlyeffective way to communicate”

All verticalsThe simple fact is, consumers prefer messaging as acommunication channel. According to Ovum, over 70per cent choose it over other media. This trend is trueacross all verticals – banking and finance, retail, socialnetworking and media, OTTs, travel andtransportation and healthcare. They all use it toacquire, retain and grow their customer engagement.

With enterprises looking to reach customers anytimeanywhere and provide the best contextual experience,there is no better way than mobile messaging. Thechannel allows for assured reach, delivery and two-way engagement.

SAP Mobile Services is now working to make availablemobile communications across all digital enterprises.Willock says: “This is extended to SAP product suitestoo as an integration for end customer value.Especially in cloud-based HR and marketing offerings,this integration enables enterprises to provide a richerand more complete experience to their end users.

“For example, an international organisation might useexisting survey tools in the market to run a campaign.That’s fine. But it would be its own little world.

“It wouldn’t integrate with the CRM systems. Youwouldn’t be able to connect the results with who thecontact people are, what are the next steps, what thebilling issues are and so on. But in this new world,they have a dashboard that brings it together. It’s allintegrated.”

Mobile v emailAnd the key point is that this is all happening onmobile, rather than the previous option for incomingcommunications: email. Willock adds: “As we all know,mobile is immediate and email is not. People can besigning contracts, approving credit notes and so on inreal time from the phone. It doesn’t sound like a bigdeal, but it makes a huge difference.”

Of course, given the strategic nature of thesecommunications, customers demand a certain levelof quality. Willock says: “There are numerous offeringsin the market with various price points and in thisspace cheapest is not always the best so enterprisesneed to understand and evaluate the right option fortheir end to end needs.”

Holistic With application-to-person (A2P) messaging on therise in the short term, offerings need to be holistic,easy to integrate and provide reliable reach.

On the enterprise side, companies need to considerthe role of messaging in their whole spectrum ofcustomer engagement from acquisition to retention.

Mike Willock, VP forEMEA Sales, SAPMobile Services

12 13

Page 8: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE TWO FACTOR AUTHENTICATION MEF FUTURE OF MESSAGING TWO FACTOR AUTHENTICATION

Despite recent scare stories, SMS pass codes remainthe safest form of two-factor authentication, says RobMalcolm, VP of marketing and online sales at CLX.

For years, two factor authentication (2FA) was seenas an essential weapon in the fight against onlinefraud. Again and again, criminals easily bypassed

passwords and PINs. Far safer then to have users enterproof that they own the device registered to the service- like a smartphone. Adding this secondary line ofprotection was a no-brainer. And the best way to do itwas with a one-time passcode sent by SMS.

But in 2016, something happened. Headlines declaredthe age of SMS 2FA to be over. It’s insecure, they said.They were reporting an announcement made by theNational Institute of Science and Technology (NIST) inthe US. It had found flaws in 2FA via SMS messages,and said it was considering these risks and may“deprecate” SMS in future standards.

The NIST was concerned that hackers could exploitflaws in the SS7 protocol that operators use to enableroaming. They could fool the network into thinking a

device is on another network and thereby allowcommunications to be intercepted.

ExaggeratedAnd it was right. Hackers can exploit this loophole. Butaccording to Rob Malcolm, VP Marketing at CLX, the riskhas been hugely over-exaggerated. He says: “Really, theonly people who could repeatedly exploit this flaw atscale would be rogue employees inside of a GSMnetwork. It’s the equivalent of a Facebook stafferaccessing your Facebook account. So, yes, it is a risk tobe taken seriously, but I think it’s been exaggerated.”

Malcolm adds that the reaction against SMS ignores thefact that it is generally safe and – most important –widely used. “We must be realistic and offer securitythat people will actually adopt,” he says.

“There’s always a balance between security and ease ofuse. In this respect 2FA with SMS is the best option we

Is thatreally you?

have. It’s fast and people know how to use it. We shouldthink hard about dismissing it when it’s so popular. Wecan’t have users going back to just user name andpassword, which they will if we take 2FA with SMSaway.”

“I think saying we should deprecate SMS for OOBauthentication because of these vulnerabilities is likesaying we should deprecate TCP/IP because of avulnerability in a firewall or in SSL. Clearly the solution isto fix the vulnerabilities, not abandoning the use SMSfor two factor authentication.“

GullibilityMalcolm has observed that as much as 20 per cent ofall A2P (application to person) messaging on his networkcomes from authentication. Banks, social networks andothers have clearly found the process effective. He addsthat many operators are now installing SS7 firewalls tomitigate against the above risk.

While it’s always possible to close loopholes in a system,it’s harder to reduce people’s tendency to be ‘taken in’.The fact is that social engineering – and not technicalhacks – is behind most attacks. In the case of 2FA withSMS, it’s criminals persuading a call centre agent todeactivate the original SIM, and provide a new one. Orphishing users with a spoof text message or email.

Malcolm laments that these attacks exist, but believesSMS two factor authentication can help. He says: “99 percent of attacks are by social engineering rather than anykind of technical hack. I’d argue that 2FA can help toreduce these attacks, because it makes them so muchmore complex than merely trying to gain someone’s username or password.”

Rob Malcolm, VP of marketing and online sales, CLX

14 15

Page 9: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

More and more companies are using mobilemessaging to improve customer engagement,make services more efficient, generateincome or save money.

In this section we highlight real use cases bytrain operators, banks, schools and more.USE CASES

0216 17

Page 10: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE TEXT DE-CONGESTION MEF FUTURE OF MESSAGING GUIDE TEXT DE-CONGESTION

Virgin Trains uses simple text alerts toreduce passenger surge at one ofBritain’s busiest stations. SteveGooder, head of digital technology atVirgin Trains, explains how...

Keeping calmat the station

The 7pm London to Glasgow Virgin train is popularto put it mildly. This is an extremely in-demandservice at all times, but at 7pm the price drops.

Inevitably, demand soars. And that means congestion ashuge numbers of people stand in front of the departurescreens waiting for the platform to be announced.

But 20 per cent of these passengers – pensioners,and those with families or disabilities – get a covert

71 million people a year pass through Euston stationin London. That’s on a par with Heathrow. Exceptsmaller. Much smaller. Steve Gooder, head of digitaltechnology at Virgin Trains, says: “We take thecongestion issue very seriously because we don’t justcare about what happens on our trains. We thinkabout the end to end journey. And if you’re caught ina crush, that’s not the best start.”

Split the packThe challenge for Virgin Trains was finding a way totackle congestion without changing the systems at thestation. Why? Because Virgin Trains doesn’t run Euston.Network Rail does. “We were dealing with legacy trainissues,” says Gooder. “We couldn’t re-engineer thestation or split the train in two or make it leave later. Sowe concluded that the best way to attack the problemwas to identify a subset of travellers and give themadvanced notice – to split the pack, so to speak. Andthe best way to do that was through text.”

Gooder and his colleagues devised a process thatintegrates with the screen announcement system.When the button is pressed to confirm the platform,there’s always a 75 second delay before it isdisplayed. At this point the system batch messagestexts to the priority group, who are identified whenthey buy their tickets online. It can also send an SMSmessage 20 minutes earlier to confirm a train isbeing prepared.

Now or neverThe immediacy of SMS made it the perfect mediumfor the project. But there were still wrinkles for Virginand its messaging partner OpenMarket. Gooder says:“We had to work around some scenarios. Forexample, if a customer is on the underground whenthe message is sent, we can’t have them receiving atext when they arrive at Euston saying the train isready when it has already left.” Happily, that has nothappened. The system won’t re-send a message if itis not opened within a few seconds.

Gooder is delighted with the results. He says thecompany’s Net Advocacy Score (which gauges theloyalty of a firm's customer relationships) has risen12 per cent among those that use the service. Andsometimes he tracks the impact in the mostunscientific way – by popping out of his office,standing at the back of the station and watching.

“The challenge wasfinding a way to tacklecongestion withoutchanging the systemsat the station.”

headstart. They head for the train 75 seconds beforethe surge begins. That headstart breaks up the traffic.Everyone benefits. And it’s all thanks to SMS.

The text project is the brainchild of Virgin Trains,which is based at Euston and uses the hub as its‘home’ station. It kicked off the service in February2016 after months of thinking about these surges onthe concourse. After all, according to Network Rail,

18 19

Page 11: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE CARD FRAUD ALERTS MEF FUTURE OF MESSAGING GUIDE CARD FRAUD ALERTS

FICO helps banks combat fraud by sendingtexts to cardholders at the moment ofpurchase. Gabriel Hopkins, senior directorof product management at FICO explainswhy it works so well...

Raising the flagon card fraud

Every time a bank flags up what it thinks mightbe a suspect transaction it has three choices:decline it (and risk annoying an honest

customer); approve it (and reward a criminal); contactthe cardholder to see if it’s genuine. The latter isclearly the best option. But for years it was notpossible. Now, thanks to smartphones, it is.

Which is very interesting to FICO. The US companyhas been combating financial fraud for 60 years.More accurately, it’s been in the business of creditscoring and analytics. The company began byapplying maths to the probability that a person couldpay back a loan. Before FICO, such decisions lay at thewhim of the bank manager. The obvious superiorityof the FICO system meant the company grew rapidly.It became an expert in using advanced analytics todetect and predict all kinds of consumer behaviours,including abnormal activity that could be fraud.

Today, FICO protects 2.5 billion card accounts, andnearly two-thirds of all credit card transactions. Itdoes this mostly via its Falcon Fraud Managersoftware, which uses artificial intelligence to helpfinancial institutions detect fraud – with minimalimpact to customers. FICO says Falcon reduceslosses from payment card fraud by up to 50 per cent.

And a key reason it can do this is the simple text alertat the point of sale. FICO added a mobile dimensionto Falcon when it acquired Adeptra, whose solutionsintegrate voice, SMS, email and other channels backinto a business’s host system. Today, its CustomerCommunication Services function sends 300,000texts a day to customers in the EU and 150,000 tocardholders in the US and Asia respectively.

Gabriel Hopkins, senior director of productmanagement at FICO, says: “Previously, banks had

desks of fraud analysts and basically they would calland ask ‘did you buy this iPad?’. Well, we automatethe hell out of that process. And recently we addedSMS as an option for contacting the cardholder. It’s afantastic channel. It helps us deal with high volumeswithout hiring more people. And, more important,customers like it.”

ConvenienceWhy do they like it? Well, partly it’s practical. Acustomer in the middle of a transaction might nottake a call from an unknown number. SMS is alsoasynchronous, so a cardholder can read the alert andrespond when convenient. FICO works withmessaging aggregator OpenMarket to ensure textsare sent (and arrive) promptly. But even when amessage is delayed (because a recipient is out ofrange, for example), it is still productive.

Hopkins explains: “Obviously, immediacy is goodwhen you’re tackling fraud as it happens. But if aperson gets a text a minute later or even 20 minuteslater, it can be reassuring. As a customer, you mighttrust a card more if you get good communication.”

Water proofAdding this extra layer of security to cardtransactions has had a huge impact. Hopkins says itsaves one bank customer tens of millions of poundsin a year. And it’s not just banks that benefit. FICO hasalso worked with Thames Water to automate voiceand text calls to customers with overdue bills. Thesecustomers can pay without ever having to speak witha collections agent. And they do. In the first sevenmonths of using the FICO system, Thames Watersent 345,000 automated voice calls and 95,000 textmessages. It collected £10 million.

Now, FICO is keeping an eye on new messagingbehaviours to ensure it reaches customers in theirchosen channels. It’s looking at chat apps, and alsopush notifications.

But it believes SMS will always have the edge. “TheOTT platforms can offer a richer experience,” saysHopkins. “There’s the scope to put in a bank logo andmore contextual information. But against that, theyare closed and mostly they require the user to initiatethe connection. The great thing about SMS is that it’sopen. No one has to set anything up. It just works.”

“Adding this extralayer of security to

card transactions hashad a huge impact. It

saves one bankcustomer tens of

millions of pounds ina year.”

20 21

Page 12: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE THE MACHINES THAT TEXT ’FIX ME’ MEF FUTURE OF MESSAGING GUIDE THE MACHINES THAT TEXT ’FIX ME’

was any change in temperature. Later he called inmessaging specialist OpenMarket to handle the routingof the texts to the engineers. He says: “We consideredemail, but the point about text is that everyone has aphone, everyone can read SMS and messages getthrough instantly. Meanwhile working with theOpenMarket API meant we didn’t have to worry aboutnetwork coverage or permissions.”

Success helpsIt took Romero six months to build the prototype, butthe tests were a success and now 150 machines acrossthe US, Germany and India have the ability to send‘help’ messages. In two years, they’ve sent thousandsresulting in that $250,000 saving.

Interestingly, though the project has kept machinesrunning, it has actually required more man hours tomaintain. However, Romero says this is deliveringunexpected benefits. He says: “With all the extra data

we have now on machine failures, we can build a moreholistic view of what works. We can start thinkingabout replacing parts that we sense may be about toexpire, for example, rather than waiting for an alertsaying they are already failing.”

Philips Healthcare is now looking to embed messagingfunctions into 3,000 machines worldwide, and isconsidering extending the concept to otherdepartments for use in nuclear medicine, cat scan andcardio vascular systems.

When hugely expensive patient imaging machines warm up unexpectedly, thedowntime costs a small fortune. John Romero, a National Support Specialist(MRI) at Philips Healthcare, explains how the problem is solved with text...

Like anyone, engineers working for PhilipsHealthcare pay special attention when texts arrivefrom a small group of contacts. Spouse, boss,

accountant maybe. But nothing will make them jumpup faster than when an SMS arrives from one of theirpatient imaging machines. Yes, you read that right. Atext from a machine.

Fix meIn 2014, Philips Healthcare began fitting theseexpensive devices with the ability to send alerts tospecified numbers. The messages essentially say:there’s something wrong, come and fix me. And thepilot has been hugely successful. The initial test with30 machines saved Philips Healthcare $250,000.

Patient imaging machines are hugely complex, medicallyessential devices. They use superconductor magnetsthat must be kept chilled at all times, and use liquidhelium as a coolant.

However, when the chiller fails, so do all the other parts.A rise in temperature will turn the liquid to gas, whichexpands rapidly and causes a cascade that shuts downthe cryogenics system. When this happens, engineersneed to restore the machine at room temperature. Suchfailures can cost up to $100,000 to resolve.

Finding faultJohn Romero, a national support specialist (MRI) atPhilips Healthcare, was thinking about this problem in2014. He says: “In the past, customers would call whena machine went down. But you can’t expect a customerto anticipate a fault. Their priority is getting the patientimage done, not the functioning of the machine, so theywould keep working till it broke down. We had to figureout a way for the machine’s thermostat itself to tell theengineers when something was wrong.”

Eventually, Romero rigged up a module using aRaspberry Pi, which triggered a text alert when there

Chill! SMS providesa cool machine fix

John Romero, national support specialist , Philips Healthcare

22 23

Page 13: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE TEXT FOR BETTER HOMEWORK MEF FUTURE OF MESSAGING GUIDE TEXT FOR BETTER HOMEWORK

A simple well-worded SMS to parents has delivered real benefits in theclassroom. Here’s the full story from the UK’s Behavioural Insights Team...

Nudging children tobetter marks – witha text to parents

What’s the best way to improve children’sschool grades? Obvious answers mightinclude iPads for every pupil. Or new

buildings. Or after-school classes. But what about asimple text to parents? A less obvious tactic perhaps,and certainly a cheaper one. But in a landmark trial inEnglish schools, the experiment improved theaverage student’s progress by a month.

Almost 16,000 students in 36 English schools tookpart in the year long trial, which was devised byHarvard Kennedy School, University of Bristol and theBehavioural Insights Team. The BIT – better knownas the Nudge Unit – was set up by the UKgovernment in 2010 to apply simple tactics based onthe psychology human behaviour to improve public

services. The aim is to make small changes butachieve large beneficial effects. The project bases itsideas on the 2008 book by Richard Thaler and CassSunstein called 'Nudge’. One of its early successescame in tax collection. When it told late taxpayersthat most people in their town had already paid,payment rates went up by 15 per cent. Thisgenerated £30 million of extra revenue a year.

So could the same methods improve pupilperformance? The researchers decided to experimentwith text alerts to parents. Raj Chande, senior advisorat BIT, explains why: “We know that parentalengagement is important, but it can be hard to reachparents who are disengaged from the education oftheir children. This is especially so in secondary

education, where the curriculum is more difficult andchildren make their own way home so there isn’t a’school gate’ culture. We looked at this and decidedtext would be the best way to reach them."

Needless to say, the wording of the texts wasparamount. The team realised that a generic requestto take more interest in their child’s schoolworkmight come across as heavy-handed. Instead, theychose a more neutral but specific approach. Theydevised around 30 texts (roughly one per week)about upcoming tests, missed homework and whattheir child had learnt that day. One example: “Yourchild has a science test coming up: make sure you askthem about cells when they get home”.

One month betterTo repeat, the trialists achieved one month of additionalprogress compared to other children. There wasreduced absenteeism too. And the scheme was cheap.The cost per pupil was £5 – equivalent to roughly 0.1per cent of the average school budget. It was also easyto implement. After all, most schools have an existingsystem for sending texts to parents usually reservedfor alerts about school closures and so on.

The project wasn’t wholly successful. Whileimprovement came in maths, English was relativelyunaffected. “I think this might be because mathslearning is all done in school, while Englishattainment is affected much more by background.We'd like to explore this further” says Chande.

Bringing wisdomUltimately, the programme was vindicated not justby results but by the attitude of participants, whobroadly welcomed the intervention. Chande says:“After a separate but similar experiment where wetexted some kids' parents about upcoming mathstests, we then asked if they wanted us to tell theirparents about their next maths test. Of the controlgroup, whose parents we did not text, 40 per centsaid yes. In the intervention group, whose parents wedid text, 55 per cent said yes.

“They said they didn’t necessarily enjoy the process,but they realised it was good for them. That reflectsa level of wisdom, I think. And one parent said thescheme brought her closer to her daughter. A prettyprofound impact for a simple text message.”

“Some children saidthey didn’t necessarilyenjoy the process, butthey realised it wasgood for them. Thatreflects a certain levelof wisdom.”

24 25

Page 14: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE SMARTER CHARITY GIVING MEF FUTURE OF MESSAGING GUIDE SMARTER CHARITY GIVING

SMS has been a big boost to charitable giving. Now, says Cancer ReseachUK’s Lisa Elkins-Jarrett, new tools are making it even more effective.

What gives? Charitiesget smart with text

It’s no exaggeration to say that SMS hastransformed the charity sector. Most people wantto give, but some can be reluctant to engage

directly with charities for fear of a ‘hard sell’. Textbypasses this. It’s straightforward and immediate.Read rates average at around 95 per cent. Forcharities, which rely so much on spontaneous actions,this is very significant.

Lisa Elkins-Jarrett, product manager for marketingplatforms (digital) at Cancer Research UK, says: “In acluttered email world, SMS gives a chance to getacross relevant and timely messages to supportersthat help increase engagement and conversion.”

Selfless selfies These factors have delivered phenomenal results,raising sums in short time that would have beenunthinkable before. In 2016, for example, CancerResearch UK asked women to post a make-up freeselfie and donate £3 using the hashtag#nomakeupselfie. They were then asked to nominatethree friends to do the same. Inside 48 hours, thecampaign raised over £2 million. Another campaign –Stand Up To Cancer – seamlessly blended SMSdonations with Gift Aid contributions. This helped tobring in more than £15 million.

Such successes have led to a huge increase incharitable text campaigns. However, the sheerpopularity of the channel has led to complications.One is obvious: more texts equal more failures. Elkins-Jarrett says: “As more charities employ SMS, higherasks become more common place. We’ve seen failurerates increase, and with it, support cases rise too.”

Typical problems include people misspelling text todonate codes, texting the wrong keyword, or reachingspend caps and running out of credit. At the height ofthe problem, Cancer Research UK found failedmessages were averaging at eight per cent andhaving a real impact on income.

Spelling failSo the charity worked with its messaging partnerOpenMarket to build more intelligence into itsservices. It installed a system to detect misspeltkeywords and send automated replies explaining theproblem. The text asked them to try again andreassured them that no money had been taken. Thisremedy corrected 20 per cent of failed messages.

The charity applied the same approach to tell asupporter had no more credit or had reached a spendcap – and then gave options on how to rectify it.

InboundAnother challenge was to apply this intelligence toinbound customer service queries. The charity foundthat, the more it used its short code to driveengagement, the more people would reply withgeneral questions. Typically, it would reply manuallyto these messages on a case by case basis. This wastime-consuming, and the charity was concerned atsending answers at inopportune times.

However, on investigating the data, it found that threetopics accounted for a third of all queries. It’s nowplanning to scan messages for key themes and thensend an immediate bounce back SMS that answersthe question.

26 27

Page 15: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE TEXT – UBER’S SECRET SAUCE MEF FUTURE OF MESSAGING GUIDE TEXT - UBER’S SECRET SAUCE

It takes a huge amount of complex technology to make Uber’stransformative service run smoothly. But without reliable text from driverto rider, it all falls down...

Text: the driverbehind the drivers

What’s behind the ‘miracle’ of Uber? Here’s acompany built by founder Travis Kalanick ona simple premise: tap a button, get a ride.

For millions of people used to waiting ages in the rainfor a taxi or struggling to find a mini-cab number tocall, it really is a kind of miracle. And they have trulyembraced it. Uber launched in 2010 and now hasquarterly bookings of more than $5 billion.

Making that magic happen requires lots of processingand hugely complex algorithms. Riders must bematched with drivers, routes plotted, paymentsallocated and shared. But there’s another essentialcomponent to the Uber success story that is lesstalked about. Text messaging.

Real time onlyUber uses SMS to keep riders up-to-date with thestatus of their request — when a driver accepts,when he or she is less than a minute away, or if theride has been cancelled for any reason. Clearly, thesemessages must arrive in real time. When they don’t,the whole process falls down. And when Uberlaunched this happened. A lot.

Kalanick says: “Uber is a pretty seamless experience.It feels like living in the future. You can’t have yourown private driver without having a great customerexperience. And if you’re waiting 10 minutes and youhaven’t been told, there’s a problem. And it’s crucialfor the rider to have that information so he canconnect with the driver. Initially, people were notgetting the high quality experience that we werepromising.”

Precious sleepUber had built its platform with a hosting provider, butdecided to go down a different route. It teamed upwith Twilio, which works specifically with app makers,letting them embed voice and text functions directlyinto their services using software.

Kalanick now says: “I sleep easier, and my engineerssleep easier because we’re not dealing with situationswhere it’s taking 15 to 20 minutes for a text to bedelivered.”

Travis Kalanick, founder of Uber

28 29

Page 16: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE SMARTER MESSAGING FOR AFRICA MEF FUTURE OF MESSAGING GUIDE SMARTER MESSAGING FOR AFRICA

30

Are Africa’s feature phone users locked out of rich media marketing? Notnecessarily. Clever new messaging formats can reach everyone, says Waheed Adam, chairman of iTouch...

For all the hype around Africa’s smartphonerevolution, the stats tell a more nuanced story.Yes, smartphones are on the up. But so are

feature phones. Indeed, according to IDC, they’regrowing faster. It says 23.1 million smartphones wereshipped in Africa during Q2 2016 (down 5.2 per centyear-on-year). But shipments of feature phones hit29.8 million – up 31.9 per cent.

Does this mean millions of feature phone users arelost to marketers, brands, utilities and government?Not necessarily. They may not be able to downloadapps or content-heavy HTML5 web pages. But there’salways messaging. The question is: can messagingever match the kind of rich experience of a native app?

Rich featuresAccording to South Africa’s messaging specialistiTouch, yes it can. iTouch has developed a platformcalled MEMS (Multi-channel Embedded MessageService) that adds rich media features to a simple textmessage. A MEMS message contains a URL thatpoints to a unique campaign-specific landing page,which includes content such as video, voice andgraphics. It can also be directed to a company’s socialmedia links and has an email function that allows thecampaign to be sent to a computer if required. All ofthis will work on an internet-enabled feature phone.

Of course, it is possible to embed a link an an ‘ordinary’text message. The difference with MEMS is that eachMEMS has a unique URL pointing to each MSISDNmaking the message personal, private, secure andtrackable. It also has interactive capabilities.

Waheed Adam, chairman of iTouch, says this opens upthe medium for all manner of use cases. He says: “This

means a bank could use it not just for sendingmarketing alerts, but also offering card or accountapplications, statements and credit limit updates.”

Adam says MEMS is affordable because it’s based onthe SMS bearer, and brands can create flash-mobisitesin minutes for a specific campaign, using existingassets. Results are trackable since each customerreceives a unique link. And marketers can send as wellas receive messages inside the platform.

Payment inside the messageiTouch says banks are testing the format, which hasalready been used successfully in Europe andScandinavia by giant retail brands. And Adam believesthe medium can go even further because it supportslinks to payment gateways. He says the platform canintegrate with local processors like PayGate andsupport in-store mobile channels like SnapScan.

“When you add easy payment to the messagingchannel, then all sorts of possibilities open up,” he says.“Imagine being able to settle a parking fineimmediately. Or pay a TV licence. That’s a big issue inSouth Africa, and yet many people are not payingbecause of inconvenience more than anything else.”

Adam adds that many brands and utilities are triallingMEMS now, but there is the issue of data charges tobe resolved. Although a MEMS campaign may only bea few kilobytes in size, this can be a hurdle in Africa,where data remains expensive for most users. Adamadmits more work is needed if brands are to make fulluse of the opportunity. “Data prices will come down,”he says, “but in the meantime we need to look at ideaslike sender-pays and reverse billing which the MNOsare preparing for.”

Messaging in Africa:it’s for everyone

“When you add easypayment to themessaging channel,then all sorts ofpossibilities open up,”

31

Page 17: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE WHEN PAYMENT IS THE MESSAGE MEF FUTURE OF MESSAGING GUIDE WHEN PAYMENT IS THE MESSAGE

32

US-based Venmo is a surprising thing: a social messaging app based aroundpayments. Actually, says Venmo’s head of product Ben Mills, it’s not surprising atall. It was designed that way...

In the US everyone knows about Venmo. In acountry that still likes to use cheques, this person-to-person payments app offers a glimpse of the

future. Venmo lets people send money to each otherinstantly. They typically use it to pay for drinks ordinner or rent. And it’s become a sensation. In 2015,it processed more than $7.5 billion in transactions. In2016, it was on course for $20 billion. Venmo haseven become a verb, a bit like Google. “I’ll Venmo you.”

But what’s less talked about is that Venmo is also asocial network. Its users don’t just pay each othersmall debts, they also use those transactions as thebasis for conversations that can be witty, flirty orprofane. Who would have thought a paymentsplatform would become a medium for messaging?

The awkward factorWell, it turns out that Venmo would. The companydesigned chat into the app from the start. Everyexchange – even the private ones – has to include anote. Ben Mills, head of product at Venmo, explainswhy: “I think we realised that payment is hard, but sois talking about payment. So we wanted to re-frameit so that it was less awkward. By adding in socialfeatures, we could make Venmo became more aboutwhat you’re doing than what you’re paying.Ultimately, if you need to pay someone $5, doing itwith a note or an emoji makes it less awkward.”

This social-by-design strategy was, of course, wildlysuccessful. Users love the emojis in particular, with –unsurprisingly – the pizza icon being the most-used.Venmo regularly updates the function too. In 2015, itadded the option to let people autocomplete with anemoji. For example, when someone types ‘rent’, theycan use a pic of a house and a wad of dollars.

Talking with stalkingHowever, other user habits emerged entirelyunplanned. ‘Venmo stalking’ is a case in point. Thisdescribes how some users will visit friends’ pages tosee and read their transactions. They can find outwhere they’ve been and who they’ve been there with.

“Obviously, this wasn’t something we created or evensaw coming,” says Mills. “But it’s a great example ofan emergent behaviour – a bit like the hashtag onTwitter – that comes when you create a community.”

Of course, the strength of Venmo’s social componentmeans that sometimes payment is secondary to theconversation. Mills says it’s common for users to paysmall sums as a pretext to a message session.“Actually, lots of payments are short conversations. Wesee lots of $1 transactions that are just an excuse tosay thanks or good job.”

Messaging makes (business) senseIf Venmo saw embedding chat as a way of removingawkwardness from the platform, it also saw a strongbusiness case for it. Mills says: “Payment is difficult,but ultimately it can be commoditised. That’s not thecase with social. People go where their friends are,and that makes social media a bigger ecosystem tomove into. And once you have a community, you cango in different directions with it.”

Those new directions include merchant payments.The aim is to help merchants go where theircustomers are - i.e. Venmo. And also to pursue thedream of social commerce. “A user might see onVenmo that a friend bought a Beyonce ticket. Thatcould inspire them to buy one too, and it will be easyto do so,” says Mills.

Venmo: turning IOU into ‘how are you?’

“By adding in chat,we could make theapp became moreabout what you’redoing.” than what

you’re paying.”Ben Mills, head of product at Venmo

33

Page 18: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE THE RETURN OF PRIVACY MEF FUTURE OF MESSAGING GUIDE THE RETURN OF PRIVACY

Mobile messaging taps into the humandesire for contact. But what about thehuman desire for privacy? Alan Duricwants to bring it back. He talks abouthis privacy-first chat app Wire...

Just betweenme and you

The success of mobile messaging proves onething. However fast technology moves, peoplebeings will always want to communicate with

other people. Every social network – Facebook,SnapChat, Line, WeChat – just reminds us of thisbasic human need. But in the rush to embrace theseexciting new forms of communication, has anotherbasic human need been overlooked? What happenedto privacy? Was Mark Zuckerberg right when he saidprivacy is ‘a social norm that has evolved over time.’?

The makers of the fast-growing messenger platformWire don’t think so. They launched the Wire app in2014 to bring privacy back to digital conversations.The web site stated: “In the physical world we talkwith each other directly. We can lower our voices orclose a door to share private thoughts. In the onlineworld we should be able to communicate withoutpassing our private communications throughcorporate data mines.” So the app launched with noads, no profiling and end-to-end encryption of voice

calling. It then added end-to-end encryption ofeverything sent over the app – text, pics, video, music.That differentiated the service from competitors suchas WhatsApp and Telegram, which offer only someparts their services in encrypted form.

For Wire co-founder and CTO Alan Duric, this isincredibly important. “Most messaging platforms werebuilt for a world that was just about text. But nowthere’s video and music and collaboration and images.And people love to share these assets. So we wantedto make sure absolutely everything is encrypted.”

Next stop: IoTBut do people really care? After all, millions seemhappy to engage with services that share their data.Duric says: “I see the privacy issue as a little like thepassive smoking debate. It’s taking a while for peopleto realise the damage it can do. But attitudes arechanging. For the moment, though, security isprobably a bigger concern.”

Protecting people from the criminals who areincreasingly lurking inside message apps is critical,says Duric. And he believes if Wire can makemessaging apps safe, it can also do the same for theIoT where the threat is arguably even greater.

He says: “We’re already reading about hackersintercepting baby monitors. And about widespreadransomware. But imagine the ransoms they coulddemand if they took control of your car?” So this iswhere Wire will go next. In 2016 Wire open sourced itsencryption protocol and its mobile, desktop and webapps. The next step is to open up the API so thirdparties can build their own secure services on top.

Road safetyDuric says: “Data in the cloud will always be prone toattacks. So the only way to really protect the IoT is toensure the data at the end points is properly encrypted.That’s what Wire can do.”

Thus, a platform originally designed to make it safe andprivate to talk to your friends could end up securing theway a car ‘talks’ to its manufacturer. “In the future a caris going to be like an iPhone. It will have so manyconnected components and occasionally it will need asoftware update. Our platform will help companies likecar manufacturers build security into these updates.”

“I see the privacy issueas a little like the passive

smoking debate. It’staking a while for people

to realise the damage itcan do. But attitudes

are changing.”

Alan Duric,CTO, Wire

34 35

Page 19: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE WECHAT INTO AFRICA MEF FUTURE OF MESSAGING GUIDE WECHAT INTO AFRICA

36

WeChat has already conquered China. Can it do thesame in South Africa?. Brett Loubser, CEO of WeChat .talks about rising to the challenge...

WeChat’sAfricanadventure

Fans of South African radio DJ Gareth Cliff have aninteresting way of engaging with their favouriteshock jock. Basically, they can ‘add’ him. In 2014,

Cliff decided to do something different with his popularshow Cliff Central. He was frustrated with broadcastingrules that limited what he could say. And he wasexasperated that he couldn’t engage more directly withhis listeners. So he migrated the show over to WeChat.

Today, WeChat users in South Africa can go to ‘addcontacts’, find ‘official accounts’, type in CliffCentral andadd the show. Thereafter, they just tap ‘listen’ to takepart in the fun.

Of course, it is also possible for fans to listen to theshow online and debate it on Twitter et al. The

difference with WeChat is that the whole experiencehappens inside one app. In 12 months, CliffCentralamassed more than 131,000 followers this way.

Fraction of the costFor Brett Loubser, CEO of WeChat Africa, CliffCentralillustrates the most compelling benefit of WeChat. It’snot an app. It’s a platform. He says: “Cliff has effectivelybuilt a fully digital radio station: the perfect showcasefor him. But he’s done it for a fraction of the cost andtime he would have spent developing an app. In fact, itwas five weeks from first meeting to going live.”

CliffCentral is helping CliffCentral popularise an app thathas, of course, come to dominate China. There, the appis so popular it effectively substitutes for the internet

itself. Over 700 million people use it to look at maps,order taxis, pay bills, find dates, get loans, play gamesand more. It has more than 10 million third party apps.

Strong traditionNeedless to say, WeChat South Africa has fewer. ButLoubser is working hard to change that. He has faithhe can. WeChat Africa has powerful backers. It is ajoint venture between Chinese owner Tencent andSouth Africa’s media giant, Naspers. And there’s astrong tradition of social messaging in South Africa,which had its own homegrown chat app, MXit, longbefore most other countries (MXit closed in 2016).

Though WeChat faces strong competition fromWhatsApp, Loubser believes its platform qualities canhelp it succeed. He says: “App development isexpensive and developers are hard to find. Our job isto show startups that they can get a service liveextremely fast and they don’t have to worry aboutpayments or location and so on. It’s all baked in.Adding any new feature is just adding an API. It’smany times cheaper and many times faster thanbuilding an app.”

Reducing the payloadAnd there are similar benefits for end users. “There’s acertain amount of payload if you want to download anapp,” adds Loubser. “There’s the data cost and findingspace on the phone. It’s all much simpler inside WeChat.”

To encourage startups to try the platform, WeChatannounced a 50 million rand ($3.5 million) seed fund,which backed micro-jobbing service M4Jam (Moneyfor Jam) and delivery service Picup among others. Italso teamed up with Standard Bank to create WeChatWallet. Users can connect their debit and credit cardsto it, and withdraw cash at ATMs by entering a code.

Bank to the futureStandard Bank also uses WeChat to offer ‘always onbanking’. Users can message an agent at any timeand receive a response inside the app. Loubser saysthis asynchronous experience works for both sides.“It’s so much better than IVR and being on hold withelevator music,” he says. “Psychologically peopledon’t mind waiting for a response, even if it takes anhour, as long as they feel in control. And the systemlets agents manage multiple queries at once, so it’smore efficient.”Brett Loubser, CEO of WeChat Africa

37

Page 20: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

The future of mobile messaging may well bepost-human.

First, there are the bots - virtual ‘people’ youcan chat with to organise your expenses,learn languages or catch up on the news.

Then there are smart machines that relayperformance information back to theirmakers, or to other machines, by text.

In this section, we highlight the pioneers in thisexciting new phase.

MACHINE MESSAGING: 03

38 39

Page 21: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE CONNECTING THE MACHINES MEF FUTURE OF MESSAGING GUIDE CONNECTING THE MACHINES

40

A new breed of connected machineswill need a new kind of SIM when theytalk to each other, says RobertGerstmann, MD of the enterprisedivision at CLX Communications...

Switchingon themachinefuture

pay for their actual usage instead of a traditionalsubscription model with fixed monthly fees. And there'sno need for the client to have individual SIM cards andagreements in every country themselves."

Gerstmann admits that operators differ in theiropenness to the idea. But he says it just reminds himof the early days of A2P SMS, when some carrierswanted to sell direct and not work with intermediaries.However, CLX triumphed in that market by anticipatingtrends and building its own platforms. Significantly, itbought Symsoft and its SMSC (short message servicecentre). That became the basis of the company'smessaging platform and gave CLX the ability toleapfrog most others in the market, who tended tohave older legacy systems.

Profitable“The fact that the SMSC was being licensed to othercustomers meant that the platform became – and stillis – a profit centre for us, while everyone else in A2Phas their platform as a cost centre. Meanwhile wefocused on one thing: the operational cost of everytransaction,” says Gerstmann.

CLX has kept growing without taking any externalfunding. And its status as a market leader was provedwhen, in May 2016, it bought 100 per cent of theshares in Mblox, arguably the creator of the A2Pmarket. The move instantly gave the Swedish firmvaluable new relationships, technical capabilities and abig bump in scale.

Now, CLX is ready to capitalise on the steady growth inA2P traffic while pursuing the IoT market. "I believe A2Pmessaging will continue to grow between 10 and 20per cent per year depending on country," saysGerstmann. In the industries where it's being used,there are still so many companies that don't use it. Andin emerging markets rising subscriber penetration isstill driving growth. In addition I'm certain there aremany more use cases that no one's thought about, sothe market is there."

What happens when you sell a connected caroverseas? Which operator does it connect toso that the driver is able to call, message and

access value added services from its dashboard? Withphones it's easy: the user just inserts his or her SIM.That's not perfect for cars. And it's even less practicalfor meters or lighting or any of the other promisedsmart devices of the future.

CLX Networks is determined to solve this problem.How? With a new breed of SIM cards that can connectto – and dynamically switch between – multipleoperators. It is currently in beta with 10 clients. Theproduct is built to make real the promise of the Internetof Things: to make previously dumb devices smart.What makes the move remarkable is that CLX is barelyeight years old. Today, CLX is NASDAQ-listed andemploys 300 people.

Cloudy prospectsThough it launched as an aggregator of A2P(application to person) messages, CLX’s plan was neverto just get better at delivering text campaigns forenterprises. Instead, it became a provider of cloud-based communication services. That led the companyto the forefront of a revolution that's puttingconnectivity inside machines.

Robert Gerstmann, MD of the Enterprise division of CLXCommunications, explains: "It's a very significantopportunity. Analysts are now forecasting 2.5 billioncellular IoT connections by 2024. Obviously, smartdevices will need a different kind of connectivity topeople. For example, I recently heard about a transportcompany with thousands of connected paymentdevices. It wanted to switch operators but couldn'tbecause switching involved getting engineers tomanually swap out all the SIMs. Having a dynamic SIMcan solve that problem."

SIMpleThere are many more use cases, of course. The systemworks well for companies whose products might besold overseas or used by consumers on foreign soil. Thealternative – keeping track on where each device is soldand fitting it with a local SIM – is clearly untenable.However, this does challenge CLX to ensure its dynamicSIMs can easily connect to multiple MNOs. Gerstmannsays: "We're working to create a global connectivityservice for the IoT market on one SIM card. Clients just

“Smart devices willneed a different kind ofconnectivity to people.”

Robert Gerstmann, MD,enterprise division, CLXCommunications

41

Page 22: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE HEAVY & SMART MACHINES MEF FUTURE OF MESSAGING GUIDE HEAVY & SMART MACHINES

Heavy industry might be the next big user of mobilemessaging, says Sethu Meenakshisundaram, presidentof SAP Mobile Services. He talks about the promise ofsmart connected machines...

You have a new message(from an oil rig)

Sethu Meenakshisundaram has a vision formobile messaging. And it’s not necessarilyhuman. He is qualified to opine on the subject.

He was a director at messaging pioneer Sybase for15 years and today he is president of SAP MobileServices (a division of SAP), which bought Sybase in2010. He’s seen messaging evolve from a person toperson medium to a channel through whichenterprises can communicate with customers. Thenext stage, he insists, is all about machines.

Supply chain reactionHe says: “There will always be human interaction,whether it’s through SMS or chat apps. But the bigopportunity as I see it lies with bringing connectivityinto fields like supply chain, diagnostics and field

management. We call it ‘intelligent interconnectthrough messaging’."

Verticals on the upThis sounds like the Internet of Things. And in a wayit is. It’s all about connecting remote and previouslydumb devices to other devices and to their humancustodians. But Meenakshisundaram is notconcerned with consumer use cases like smarttoasters. His focus is on bringing connectivity toverticals such as oil and gas, automotive, industrialtools and so on.

He says: “SAP started in manufacturing. It’s stillnumber one in transport, logistics, supply chain. Thisis a great opportunity to add value and bring new

efficiencies to some of our biggest and oldestcustomers. They can use our solutions to find outhow their machines are functioning at any time, anddo it all from one central location.”

Out to seaImportantly, thanks to the core strengths of SAPMobile Services, this connectivity goes through theSMS channel. Meenakshisundaram says this is safeand promises unrivalled reach.

“We have customers who have equipment that’s600 miles out to sea. There’s no wifi. No cable. Sowe provide them with SMS connectivity. But theimportant part is that they don’t know it as SMS. Weput it inside a beautiful wrapper that’s much morerelevant to want they’re doing,” he says.

Cellular and universalAnother critical component of the SAP MobileServices proposition is its ability to offer universalcellular connections, no matter where the deviceends up. “It’s funny. No one really thinks about thereality of connectivity,” he says.

"They want to put a diagnostic solution insideexpensive and important devices. But they don’tthink about how it will connect to differentoperators around the world. We solve that withglobal SIMs. It means that, for the manufacturer,they can use the same part number wherever amachine goes.”

Pull hereUsing SMS also brings with it a greater level ofsecurity. In these early days of the IoT most devicesare connected over IP. That has led to numerous,regrettable hacking scares. The SAP Mobile Servicessolution runs over a private and secure SS7 channel.What’s more, in most cases, the devices are notactively sending data back to base. Instead, the datais pulled by the monitoring organisation.

Meenakshisundaram explains: “It’s very importantthat these solutions are extensible andprogrammable. We build them so that the customerhas control over when the data is sent, and whopulls it. The devices are never permanentlyconnected. They’re essentially in sleep mode, whichalso extends battery life over several years.”Sethu Meenakshisundaram, president of SAP Mobile Services

42 43

Page 23: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE ACCOUNTANCY BOT MEF FUTURE OF MESSAGING GUIDE ACCOUNTANCY BOT

Every day, millions of small to medium sizedbusiness use specialist accounting software tomanage their finances, organise payroll and log

expenses. Three million of them use Sage, one of themost popular packages available. And the rest? Anexcel spreadsheet. Or, failing that, a big old shoe boxfilled with scribbled notes and receipts.

These users could be viewed as a businessopportunity for a company like Sage, which estimates65 per cent of all businesses use these ‘analogue’methods. But the truth is, most of the shoe boxbrigade are turned off by accounting software, nomatter how affordable and intuitive. They wouldrather ignore it and do more enjoyable things – likechat on Facebook Messenger. However, theseentrepreneurs do need to manage their accounts. Sowhat’s the answer? For Sage, it is simple: makeaccounts something you can do by chatting.

British accounting humourIn July 2016 Sage launched Pegg, a chat bot that letssmall business owners enter and track informationsuch as money owed, account balances, expenses

and more. Customers simply add Pegg inside Slackor Facebook Messenger (or use it inside the existingSage software) and then ask it questions in naturalconversation. They could ask ‘who owes me money?’or ‘how much did I spend on coffee in April’ or ‘whatdid I earn last month?’ and Pegg will deliver theanswers in real time and without jargon. Indeed, Peggeven uses what has been described as ‘Britishaccounting humour’. Ask Pegg what it’s doing, and itwill reply: “Daydreaming about purchase ledgers.”

Non-threateningFor Kriti Sharma, VP of bots and artificial intelligence atSage, Pegg is all about giving people a non-threateningway in to personal accounting. “Let’s be honest,accounting is nobody’s favourite task,” she says. “It’s apain point for most people. And for them the traditionalinterfaces are not working. We looked at the marketand saw how popular messaging apps have become.People love live chat. So it became obvious that chatwas the best interface to reach them.”

From the start, Sharma and her team were careful toposition Pegg not as a chatbot for accounting, but

more of a personal assistant. Hence the non-threatening name. And by putting Pegg inside Slackor Facebook Messenger, it removed further barriersto adoption. She says: “Pegg is free, and there’s nologin or registration. You just add it to your contactsand start talking. The best feedback we’ve had hasbeen from users who said they felt that accountinghad become invisible. They were just chatting.”

Personal accountancy trainingToday, Pegg has users in 85 countries (thoughavailable only in English). And it’s evolving. Forexample, following feedback from users, Sage addedemojis shortly after launch. Sharma says she wantsto see Pegg grow from being a tool to a genuine‘personal trainer’ that actively assists users inaccounting matters. She says: “The more you usePegg the more it can understand your behaviour,automate tasks and start to make suggestions. It’s abit like a fitness tracker. You might start by justlooking at the data, but in time you can quantify yourperformance.” Sharma believes Pegg represents morethan just an alternative to more formal ways tomanage accounts. She insists it can actually help new

businesses survive. According to Sage data, around400,000 businesses launch in the UK alone everyyear, but half close within the first few years.

The basicsSharma believes the administrative burden is a keyfactor in why they fail. Meanwhile, in developingmarkets, there’s an even more basic need for somekind of digital accounts management. She says: “Noone starts a business to file expenses and do taxreturns. But these things need to be done. Peggaddresses this – especially in new markets wheremobile is the main channel for people.”

Kriti Sharma, VP ofbots and AI, Sage

44 45

Chatified accountantThree quarters of small businesses use Excel or even paper to manage theiraccounts. Now, they can chat to a bot called Pegg. Kriti Sharma, VP of bots andAI at Sage, explains why the accountancy software specialist created him...

Page 24: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE TEACHERBOTS MEF FUTURE OF MESSAGING GUIDE TEACHERBOTS

Every day Duolingo helps 150 million people learn languages for free. But untilrecently, says Duolingo’s Gina Gittholf, learners couldn’t practice conversation.Now, they can – thanks to message bots...

The mere idea of a chat bot is hard for manypeople to understand. But one that speaks in aforeign language? Surely that’s just a step too far.

Not for the makers of language-learning app Duolingo.For them, friendly bots like Chef Roberto (the pizzamaker) and Renée the driver have solved a hugeproblem: how to get people talking in a new languagewithout self-consciousness.

Duolingo is a free app that helps people learn foreignlanguages through memory puzzles. Users can earnvirtual currency called Lingots, which they can spendon extras. There’s also a progress bar on top thatexpands or shrinks to reflect performance. Learnerscan also compare their progress with that of theirfriends. Essentially it employs the techniques of videogames to make learning and retaining information fun.

The chat gapThe fact that it works – and that it’s free – has helpedDuolingo gather 150 million users. But for all itsprogress, Duolingo had a sticking point: users couldn’tchat. Gina Gotthilf, VP of growth at Duolingo, explains:“One of the main reasons people learn languages is tohave conversations in that language. They want totravel and have a more authentic experience, makefriends, get a job or communicate with family. Yet theway we learn languages – memorising vocabulary andlearning grammar rules – is very different from comingup with things to say in real-time.”

The obvious solution was to pair, say, a Germanlearning Italian with an Italian learning German. Butwhen Duolingo tried this it found that people weretoo embarrassed for it to be effective. The gap inability between the two was too great. “Speaking alanguage and being able to help someone else, orteach it, are very different things,” says Gotthilf. “Also,

it's awkward speaking with strangers on camera,especially in a language you don't really know.”

Then came Roberto and the other bots. They solvethe problems arising from human encounters. In achat session, they text simple questions that demandobvious answers. Tricky words are highlighted (userscan click to see translations) and there’s anautocomplete function that kicks in when learnersare using the correct word. Chat sessions are short.

End of awkwardGina Gotthilf says: “Talking to a bot isn't awkward ordaunting at all. It's an experience that's fully in yourcontrol and provides a way to practice conversationprivately, without the fear of sounding unintelligent.They also actively help you respond when you'restuck, teach you different ways to respond to differentsituations, and let you think before you have to replyinstead of demanding an immediate reaction.”

Interestingly, these teacher-bots don’t seem to beaffected by the usual shortcomings of AI chat. In acustomer care scenario, for example, bots canmisconstrue a question or its nuance. But here, thenature of the conversation is limited and basic. And,indeed, it’s the bot that’s teaching the human.

For now bots are available for English speakerslearning Spanish, French and German, but moreoptions are coming soon. Even more significantly,Duolingo will add voice recognition so that learnerscan talk to the bots.

The bots that say Ja, Si and Oui

“Talking to a bot isn'tawkward . It's fully inyour control andprovides a way topractice conversationwithout the fear ofsounding unintelligent.”

Gina Gotthilf, VP ofgrowth at Duolingo

46 47

Page 25: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE ROBOTS ARE YOUR MASTERCARD

For decades consumer banking worked thesame way. Account holders would walk into abuilding and tell cashiers what they wanted. On

the plus side, it was nice to talk to a human. On theminus side, it could be a huge time killer. So manyhours spent travelling to the branch, queuing at thebranch and waiting for the instructions to actuallytake effect.

Digital wiped out the waiting. Banking went from thebranch to the pocket. A balance inquiry would takeseconds. A transfer could be carried out in minutes.Great. But something was lost in the transition. Digitalmade financial management a self-service affair.People had to master passwords, PINs, menus andicons. For many, this was confusing and impersonal.

Best of bot worldsNow, a third way is coming, which combines thespeed and convenience of digital with a welcomepersonal touch: banking by bot. The innovation letspeople chat with artificially intelligent agents insidea messenger app. Once they have linked theiraccounts to their messenger ID, the bot can answerall manner of questions. And for the account holder,it feels as natural as texting a friend.

Mastercard is at the forefront of this change. Itunveiled Facebook Messenger bot pilots for issuerbanks and merchants in October 2016. Eventually, itwill roll them out to Messenger’s one billion users.And the company also has plans to develop bots forother chat apps as well as SMS.

Kiki Del Valle, SVP of Commerce for Every Device atMastercard, says: “Our starting point is thatcustomers should pay how they want – and that weneed to go where they are. At the moment, they lovemessaging apps. And these apps are all preparing tosupport payments and shopping. We need to beready for this.” Once a user has selected theMastercard assistant from the bot menu and linkedan account, he or she can ask questions aboutspending, review purchase history, make transfers,monitor spending levels, set up alerts and more.

Layers of detailThe interactions can be very detailed. Del Valle says:

“You can ask the bot how much you have spent inrestaurants in the last three months. You could thenset a cap on spending and an alert for when you nearthe limit. It’s so much quicker and more efficientwhen you do this in a natural language chat session.”

However, the bots don’t just make interactions moreintuitive. They also help both Mastercard and itsissuing partners to showcase products and services.Del Valle says: “We can look at the questions peopleask to determine how our partners could do thingsmore efficiently and make the most of businessopportunities. And in the case of Mastercard we canuse the bots to tell people about purchase protection,fraud alerts or contextual offers.”

Shop a lot by botAlongside developing its own bots, Mastercard hasbuilt a Bot Commerce API that merchants can embedinside their bots to make buying with Mastercard aseamless experience. The idea is that consumers willshop on chat platforms and check out with theMasterpass digital wallet.

“It’s a safe transaction experience that takes just oneclick,” says Del Valle. “And it can all happen withoutthe customer having to leave the chat session and goto another payment screen on a web site. We thinkthis could very important to merchants as commercemoves inside messenger apps.”

Kiki Del Valle, SVP, Commerce for Every Device, Mastercard

How can AI be of assistance?

After years of self-service digital banking, the personal assistant is making a comeback. But this time the helpful agent is a bot.

Kiki Del Valle, Sr VP for Commerce for Every Device at Mastercard,looks ahead to a revolution in financial management...

MEF FUTURE OF MESSAGING GUIDE ROBOTS ARE YOUR MASTERCARD

48 49

Page 26: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE QUARTZ: THE NEWS IN TEXT MEF FUTURE OF MESSAGING GUIDE QUARTZ: THE NEWS IN TEXT

Zach Seward, senior VPof product and executiveeditor, Quartz

Can you have a conversation with the news? Digital newspaper Quartz thinks so.Its app dispenses with the usual menus and lists. Instead, it lets readers browsethe day’s headlines in a chat interface. Zach Seward, Quartz’s senior VP orproduct and executive editor, explains the thinking behind a left-field idea...

It’s a truism in tech that when a medium embracesinnovation, it takes a while for old norms todisappear. It’s why the first movies looked like filmed

stage plays and mobile phones had qwerty keyboards.In a similar vein, online newspapers tended to look backat their paper editions for inspiration. They wrote digitalnews stories to the same length and style as before.And they presented in sections – politics, sport,weather – in the expectation that readers wouldbrowse them as they did the morning edition.

But newer ‘digital native’ entrants have torn up therulebook to think differently about how we consumethe news. They include US-based Quartz, whichlaunched in 2012 to publish ‘intelligent journalism, builtprimarily for tablets and mobile phones.” It currentlyhas around 19 million unique monthly visitors.

News as APIIn 2015, Quartz dared to ask: can we have aconversation with the news? Its app, which launchedin February 2016, lets readers discover stories in a chatformat. The bot presents a story to which they can tapa caption for more or ‘anything else’ to move on. Theexperience is enlivened with gifs, videos and emojis.

According to Zach Seward, Quartz's senior VP ofproduct and executive editor, the app is the result of amission to find out what it would look like if “journalismwas truly made for peoples’ phones.” Indeed, Sewardeven likens Quartz to an API that can go anywhere itsreaders are, in whatever form is appropriate.

Thus, Quartz has a desktop site, a mobile site and anemail daily brief. But for four years, it didn’t have a nativeapp. Truth is, it just didn’t like them. “We weren’t keen

on native apps for news. And we still feel that way,” saysSeward. “Its one thing to make the web work well onmobile, but another to build a truly native mobileexperience. We considered building an experienceentirely around notifications, but that raises thequestion: then what?”

The right toneAnd so Quartz began thinking about a conversationalinterface. If millions of people were communicating thisway with their friends, why not the news? The teamstarted building. It designed the UI to be an almostidentical replica of the iPhone's stock messaging app.

But it soon became clear something was wrong. Sewardsays: “When we prototyped it, we were using existingcopy from the site and it felt lifeless. We obviously needto write specifically for the format to get thatconversational tone right.” This is important because allthe content in the app is written by staffers. It’s only theusers’ interactions that are controlled by bots.

Admittedly limitedThe Quartz app presents around 20 stories a day.Obviously, this is much more limited than a regular newsapp with menus and lists. But that’s the point, saysSeward. “The app is not especially efficient in theclassical sense. You can scan a list if you want a broadview of the news. That’s not our goal: we want to beinformative and entertaining.”

For this reason perhaps, the app has divided readerssharply. Some hate it. Others adore it. Quartz initiallyexpected readers to open the app once a day, but foundmost open it twice a day and spend between four andfive minutes with it.

The news followsthis message

50 51

Page 27: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE SLACKBOTS - MAKING WORK BETTER MEF FUTURE OF MESSAGING GUIDE SLACKBOTS - MAKING WORK BETTER

When Slack introduced bots to its workplace comms tool, they were an instanthit. James Sherrett, senior manager of accounts for Slack EMEA., explains why...

Slack’s corporate motto is ‘be less busy’. Themessaging company has found huge successmaking workplace communication easier for its

four million users. Its own stats say deploying Slackleads to 49 per cent less email use and 25 per centfewer meetings. This is partly because someone elseis doing so much of the work. More accurately,something else. Bots.

Today, there’s a Slack bot for pretty much everything.Most bots exist to automate menial tasks. Birdie, forexample, will take away the pain of managing andpreparing expenses. Workers send it privatemessages with amounts, and upload scans ofreceipts. The bot (called Bill of course) will organiseeverything inside a Google Drive spreadsheet.Meanwhile Howdy is a bot for team management. Itcan be programmed to ask questions such as ‘whatare you working on?’ to designated colleagues, thencollate and distribute updates.

HumanoidAnd to make the UI easy, Slack even gives bots‘human’ qualities. They get their own profile pics,names and bios. And they sit inside the teamdirectory, where they can be messaged like anycolleague. James Sherrett, senior manager ofaccounts for Slack EMEA, says: “A simple way to thinkabout bots is in terms of email. There are four kindsof email. There’s spam. There’s someone asking youfor information. There’s someone giving youinformation or sharing a file. And there’s a servicegiving you a notification like ‘you have a new follower.’When you’re inside Slack, bots do the last one.”

The time-saving ability of Slack bots helps to explainthe meteoric rise of the company since it launched in2014. Slack succeeded because it recognised that

messaging in the workplace was broken. And it did soalmost by accident. The firm started out as a gamedeveloper and was so frustrated by email that it builtits own in-house messaging tool. In time, thefounders realised they had built something special.But it wasn’t the game product. It was the messagingtool. Sherrett says: “When the game closed, the teamall said they never wanted to work in an organisationwithout the messaging system, so they pivoted. Andwhen 8,000 companies signed up for a trial inside 24hours, it was clear they were on to something.”

Inbox imprisonedSlack’s early research revealed that while a smallnumber of its clients switched to Slack from otherspecialist workplace tools, the vast majority wereusing a mishmash of email, apps and (probably)shouting. Sherrett says: “The situation was so bad.Take email. It works for external comms, but withinthe company it’s mostly terrible. All the value islocked up inside people’s own inboxes.”

So Slack devised features designed for the reality ofteamwork. Users can chat one-on-one or in groups.They can upload and share files, and they can do soacross multiple devices including mobile. Crucially,Slack also integrates with other apps and servicessuch as Skype or Dropbox or even Uber. There are625 such apps now. And to make it even easier tocommunicate with them Slack introduced bots.Essentially, bots let users talk (via a chat session) totheir favourite services inside Slack. No need to opena separate app or worry about drop-downs menus.

Hard-working Slackbots may be making it easier foremployees to do their jobs, but they’re not all work.There are bots for Poker, Ping Pong and even Connect4. Something to do when you’re less busy.

Office chat Slack bots make work smarter

“Bots let users talk to theirfavourite services inside Slack. No

need to open a separate app orworry about drop-down menus.”

James Sherrett, senior manager of accounts for Slack EMEA52 53

Page 28: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

What part will messaging play in our digitallives? Is it possible that messaging willreplace apps and web pages?

What are the prospects for ‘conversationalcommerce’? Could chat be the medium inwhich we shop online?

And how will people send messages in apost-screen future of wearables and AI?

In this section, we explore where next forthe medium itself.

WHAT’S NEXT...04

54 55

Page 29: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

56

Benedict Evans has become mildly famous inmobile circles for his astute observations abouttechnology trends. The partner at VC

Andreessen Horowitz has made plenty of insightfulstatements. But few are quoted more than this one:“Old: all software expands until it includes messaging.New: all messaging expands until it includes software.”

In other words, messaging services are where peoplego not just to chat, but to find information, look atmaps, buy stuff. Clearly, they’re doing that already inWeChat, Line, Viber and the rest. The question is: dousers need to go anywhere else? To quote anotherEvans’ blog entry: “Can messaging become the thirdruntime on the phone, after the web and native apps?”

Evans believes the reason why messaging has acquiredsuch significance now is due to the nature of thesmartphone itself. After all, people could use MSN andSkype on the PC, but these services never achieved anystrategic dominance. He says “For two decades we hadthe monolith of mouse, keyboard and web with Googlelayered on top. But mobile has disrupted all that. In fact,the smartphone itself has become a social platform.”

He says this is because every app can access youraddress book and photo library, so adding a new app ispainless. Also, the phone number itself acts as a uniquesocial identifier which makes it easy to find (and trust)your contacts. The ease with which people can try newmessaging services (and quickly add friends) has led toan explosion of apps. And there have been hugesuccesses such as SnapChat, Kik, Instagram, FacebookMessenger and WhatsApp.

Behavioural changeThe challenge for new entrants is to find a newbehaviour that can make their offering go viral. Somehave done it – Snapchat most obviously – while others,like Whisper and Secret, could not sustain interest.

Now, argues Evans, the question for messaging giantslike Facebook is whether they can take the next stepand become platforms. The reason this is appealing isbecause it offers a way to overcome the closed natureof the phone. Evans explains: “If you want to distributeapps. You can’t. There’s no systematic way to distributeapps outside of an app store. If you don’t own the OS,you’re locked out. It’s a big strategic problem.”

In Asia, the messaging app has become a container forweb apps. As a result, users can do everything insidethem – shopping, banking, gaming and so on. And thishas turned services like WeChat into miniature versionsof the web itself. In the West, Facebook is doing its bestto become a destination for brands’ notifications.Websites can send messages like ‘your item has beendispatched’ directly into Messenger, if the user loggedinto that website with Facebook when ordering.

ConversationalAnd then there are the chat bots. Evans can see thelogic of ‘chatting’ with a brand. He says: “A company likeAmazon could send a notification to a user by email orvia Facebook Messenger, and the customer wouldprobably find FM better. It’s easier to reply to thatmessage and start a conversation. It feels like SMS.”

But he believes the fundamental problem with bots lieswith the limitations of their NLP (neuro-linguisticprogramming). “In an e-commerce transaction, botsmight work well,” he says. “There are a limited numberof things you’re likely to ask: when is my package beingdelivered, for example. NLP will guess that, and there’s

always a user agent to fall back on. But a CNN bot? Howeasy would it be for a user to break it? That’s harder topredict. That said, when bots do work, they becomevery interesting, particularly because they get aroundthe app-installation problem.”

Benedict Evans, partner at Andreessen Horowitz

MEF FUTURE OF MESSAGING GUIDE THE NEW RUNTIME MEF FUTURE OF MESSAGING GUIDE THE NEW RUNTIME

First people lived their digital lives inside webpages. Then came the app. Could messagingproducts be next? Benedict Evans, partner atAndreessen Horowitz, has been consideringthe question...

Platformthree?

57

Page 30: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE NETWORK SMARTS

Amobile operator knows a lot about its customersjust by analysing their behaviour on the network.Some of this is obvious – location, browsing

habits and so on. But it’s been revealed that operatorscan even tell when people are bored. In 2015,researchers at Telefonica developed an algorithm toreveal boredom. In a project called “When Attention isnot Scarce - Detecting Boredom from Mobile Usage”,they looked at the time duration since the subscriber lasthad a call or text, the time of day, and even things likescrolling. A series of speedy swipes through a Facebookpage was indicative of restlessness, for example.

The researchers concluded that this algorithm couldpredict boredom up to 83 per cent of the time. Theytested this by sending users a link to an article onBuzzFeed. They found that more people clickedthrough from a ‘bored’ sample than a control group.

So what? Well, Telefonica is now considering how tobuild this functionality into an wider offering for itsenterprise customers. In theory any brand could sendexisting customers a well-timed alert at periods of‘peak boredom’ for better engagement.

For Telefonica, this is just one quirky example of a driveto make more of its network assets – and thereby givemore value back to corporate messaging users.

Indeed, in late 2016 it launched LUCA, a dedicated BigData services unit, just for this purpose. “Data is acritical asset for the future of Telefónica… and we wantto help our clients understand its full potential,” saidChema Alonso, chief data officer of Telefónica. “Westrongly believe it will help our clients in decisionmaking, more efficient resource management and inreturning the benefits of this wealth of information notonly to their clients and direct users, but also to society.”

Crime stopperRelieving boredom may be at the lighter end of big dataproblem solving. But elsewhere, Telefonica is alreadytackling much more serious issues such as combatingcrime. In fact it is already using its data assets in the UKand Peru to help banks detect fraudulent activity.

The platform uses network usage data to help banksverify they are sending messages with sensitivefinancial content to the right person at the right time.

The operator SIM is a treasure trove of information abouta subscriber. Now, Telefonica is using this insight to makemessaging even smarter...

The power of the network

MEF FUTURE OF MESSAGING GUIDE NETWORK SMARTS

Using these insights, Telefonica can disclose if thatrequest has come from a recently acquired SIM, forexample, which may indicate criminality.

When in roamSimilarly, Telefonica can detect when a phone isoverseas and alert the bank. This means it will not benecessary for the customer to make a call to havepurchases approved. "The challenge for financialinstitutions is to avoid scams and frauds in the use ofcredit cards, both in the physical and virtual worlds. Thegoal is to provide the customer with a secure, seamlessand integrated experience, "says Ethel Bazán, managerof Big Data at Telefónica Grandes Empresas.

Wider verticalsIn time, Telefonica will apply its network insights tomessaging across other verticals. It believes the ideahas applications in insurance, marketing and healthcareand could help with services like travel alerts and creditscoring. Naturally, all activity will require user consent.

Interesting that this big data push by Telefonica centreson a decades-old medium like SMS. But as has been

said many times before, text is universal and doesn’trequire data access. Typically, 90 per cent of texts areread in minutes.

InflexionIndeed, Telefonica says the market for enterprise SMShas doubled in the last two years. It believes its plansto use of network insight will accelerate this growth.

James Lasbrey, global head of messaging at Telefónica,says: “We’re hitting an inflexion point now in terms ofon-to-many business messaging. The potential ishuge, and with the addition of big data to makemessages more targeted and smart, I think we’ll seesome fantastic new use cases very soon.”

James Lasbrey, global head of messaging, Telefónica

58 59

Page 31: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE IN-APP MESSAGING MEF FUTURE OF MESSAGING GUIDE IN-APP MESSAGING

What if every brand could have their own in-app messaging channel rather thanrely on WhatsApp and Facebook Messenger? Layer says they can...

An OTT that’s just for me

In the last few years, it’s become very easy fordevelopers to build previously complex servicesinto their apps with as few lines of code.

Payments? Don’t worry about all those issuers andprocessors. Just use Stripe. Location? Just add theGoogle Maps API. But what about group messaging?Should app makers just accept the need to set upWhatsApp accounts and meet customers there?

Layer doesn’t think so. In 2013, it launched a platformto let developers easily create their own richmessaging services inside their own apps. Theseservices do everything that Facebook Messenger, Kiket al can do: one-on-one and group chats, syncing

across devices, push notifications, gifs, read receipts,etc. But they allow brands to control the wholeprocess while keeping customers inside their apps.Layer now has 400 customers.

Ron Palmeri, CEO of Layer, says he founded thecompany when he began to realise the strategicimportance of one-to-one conversation. “There’s apattern that always seems to emerge. A servicebecomes successful and a community dynamicgrows around it. Soon these small and focusedcommunities become important and you don’t wantto lose them. It’s why YouTube recently let visitorshave one to one chat inside the site. YouTube

obviously saw people leaving the site to have theseconversations, and they couldn’t afford to let thathappen.” Palmeri has also observed the trust‘dividend’ of having rich customer communications.He says: “Brands are finding that great interactionleads to more trust, which ultimately leads to moretransactions. AirBnB, for example, has said that hostratings depend as much on the quality of theinteraction as the accommodation itself.”

Trunk callingThen, of course, there’s the fact that people respondto mobile messaging faster than to email. And theyoften prefer it to voice too. Palmeri cites an exampleof this from its customer Trunk Club, a virtualpersonal stylist that sends outfits to people’s homes.

He says: “Trunk Club was finding that only 33 per centof its sign ups actually converted to active customers.It’s because most people don’t really want to talk onthe phone about their dress sense. When Trunk Clubstarted to offer an iMessage type experience insidethe app, conversions went up to 87 per cent, andrevenue rose by 50 per cent.”

Palmeri attributes this to the asychronous nature ofmessaging; people can respond at their leisure. Andthen there’s the richness of the experience itself.Users can look at product images inside the sessionand swipe yes or no to them. Anything they buy issaved as an attachment that serves as a record of theorder, then the delivery and then the invoice.

Apps forever“These attachments are like a aggregated record ofall conversations. There’s no need to leave theconversation to access this record. It’s incrediblypowerful,” says Palmeri. Of course, Layer is basing itsmission on the assumption that brands will continueto have apps. In the last two years, some techthinkers have begun to question this. They say thatapps and sites could go away as brands just becomechat bots inside Facebook Messenger et al. Palmeridisagrees. But he’s happy for the notion to persist.

“The idea that big companies won’t have their ownproperties for talking to their own customers issimply ridiculous,” he’s says. “But I hope they keepsaying it. It sounds so absurd. And that absurdity isperfect for us.”

Layer’s in-app messaging hasworked well for the digitalpersonal stylist Trunk Club

Ron Palmeri, CEO of Layer

60 61

Page 32: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE IN CONVERSATION WITH OPERATOR MEF FUTURE OF MESSAGING GUIDE IN CONVERSATION WITH OPERATOR

US start-up Operator thinks it’s found a third way between the humanwarmth of high street shopping and the addictive convenience of e-commerce. And it all revolves around simple mobile messaging.

In the beginning there was trade. A person withsomething to sell met a person who wanted to buy.It was nice. It was social. You could haggle. But it

was also slow and, well, the range was limited. Thencame commerce: high street stores and, eventually,online shopping. It was convenient. The inventory wasclose to infinite. But we lost that human touch.

Is there a third way? The US start-up Operator thinksso. It says its conversational commerce platformcombines the warmth of human trade with the scaleand immediacy of e-commerce. It’s shopping bymessaging. Indeed Operator’s mantra is “messagingwill be the atomic unit of modern commerce”.

Operator launched in 2015 with a hugely ambitiousmission: to ‘unlock the 90 per cent of commercethat’s not on the Internet’. Users simply download theOperator app and then type in any kind of shoppingquery. The app’s human operators then makerecommendations, take payment and arrangedelivery – all via a chat session inside the app.

Faustian bargainJarah Euston, head of growth at Operator, says: “It’sa new personal way to shop – a return to two peopletrading. We had big box retailing, but it was a faustianbargain. We abstracted out human interaction forefficiency and infinite inventory. So we’re bringingback the human element but keeping the scale.”

Operator recognises that its platform is not reallysuited to everyday items, so it is focused instead on

purchases that require ‘style, taste and expertise. Forthis reason, it routes queries to independent expertsin areas like shoes, jewellery and furniture. Thecompany says it selects its specialists in videointerviews that test their knowledge and personablequalities. These experts earn commission on sales.

Euston says: “We’re riding two big structural shifts.There’s the supply side, with all the benefits digitalcan bring to inventory and delivery. But we’re alsoriding the wave of the knowledge economy in a newway. What’s gone before in gig economy – with Uberand Postmates and Instacart and so on – has centredon non-knowledge workers. We’re evolving this frombraun to brain. Our specialists can show off theirskills on the platform while building a business off it.”

EmpathyOperator’s focus on high-end items and humanexperts would seem to exclude bots. Actually, thecompany has begun using them inside its FacebookMessenger service to automate some of the moremundane parts of the process. Euston says: “Botscan’t do empathy. They can’t know you’re into mid-century modern furniture, but they can discern yoursize, age, gender and store that so an expert cananswer your question more quickly.”

Operator says its average purchase is $90 and that10 to 15 per cent of queries are converting. AndEuston confirmed that the service had sold 250engagement rings. It doesn’t get much more personalthan that.

“Messaging will bethe atomic unit ofmodern commerce”

62 63

Page 33: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE PAY WHEN YOU’RE CHATTING

because of the integration with social media.Somehow Venmo made payments social, and it reallystuck. This bodes well for Facebook and the rest. Ifyou can send money as easily as you can send a pic,then the adoption seems inevitable.” And clearly,when that happens, the next stage could be turningthe whole payment process into a conversation.

ForensicsYes, we’re talking chat bots. And Birch is a believer.“Online chat is already a much preferable way to talkto brands and retailers. Better than voice. It’s not justquicker, but it provides a forensic record of youractivity, which is important.

“Its inevitable that this will move to mobile apps andmessaging. Parking is an obvious case. Instead ofdownloading multiple apps all of which ask you tosign in and enter card credentials, you would just starta chat in WhatsApp and type ‘I’m parking my Volvo forfour hours’.

“You don’t even need to give your name or car or evenlocation. WhatsApp knows all this already. That’s a

compelling user experience for people and brands.”And it’s safer too thanks to the unique identity of thedevice. Birch says: “Chat with a brand by email or weband a hacker only needs your password to pretend tobe you. But with a mobile messaging app, they wouldneed to steal your phone and get through the PIN ortouch ID. It’s a lot more secure.”

Dave Birch, director of Consult Hyperion

MEF FUTURE OF MESSAGING GUIDE PAY WHEN YOU’RE CHATTING

The best payments experiences are invisible,argues Dave Birch, director of digital paymentsspecialist Consult Hyperion, He believes theywill disappear inside OTT messaging apps...

When Apple Pay launched, much fuss wasmade about its contactless paymentfeature. Look, you can pay with your phone.

No plastic! But two years on, ‘tap with Apple Pay’ hasmade relatively little impact. In fact, 2016 researchby First Annapolis said just 15 per cent of those withcompatible phones regularly use it.

This doesn’t surprise Dave Birch, director of digitalpayments consultancy Consult Hyperion. Right fromthe start he argued that contactless Apple Pay was aside show. But that doesn’t mean he’s bearish on theservice. Quite the contrary. He thinks Apple Pay andsimilar services like Android Pay will change the waywe transact. And the key to this revolution will be in-app payments.

Blurred lines“Payment inside apps is fast and seamless. Peopleare used to it, and they like it,” he says. “I thinkeventually we’ll see the lines blurring between payingin-store and paying online. You’ll be at thesupermarket and the app will wake up, know whereyou are and – when it’s time to pay – offer you a

choice of payment options and the chance to redeemall your loyalty points and so on. I think this isinevitable.” So, how does this relate to messaging?Well, as chat apps increasingly dominate the timespent online, there’s a good chance that thesepayments could be made inside messaging services.“I realise people don’t want to download an app forevery brand they do business with. So I think most ofthe activity will be within a handful of key apps, andmessaging apps will be among them.”

Social mattersOf course, this is already happening in China wherebarcode-enabled payments by WeChat arecommonplace. In the West services like Venmo andFacebook Messenger are rolling out merchantpayments. Birch believes their potential is not justdue to the fact that people spend time inside theseapps. It’s also to do with the very nature of social.

He says: “If you look at the Venmo and Pingit byBarclays, for example, you can see that actually Pingitis a much better product. And yet Venmo in the onethat’s really gone mass market. I think that has to be

Say hithen buy

64 65

Page 34: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

Is SMS due an upgrade? The world’s operators - and Google - think so. Theyhave teamed up to work on a new version of standard text called RCS (RichCommunications Services)...

SMS is many things. Many wonderful things: it’suniversal (on every handset); it requires no usermanual (everyone gets it); it works without a data

connection. These attributes make text the perfectmedium for non-nonsense communications, and indicatewhy business to consumer messaging is exploding.

The pull of OTTBut the basic functionality that makes SMS soappealing is not enough for everyone. Which is whyhundreds of millions of consumers now use so-called‘over the top’ apps like WeChat and FacebookMessenger. They offer group chat, messaging historyand the ability to share multimedia content and see ifa recipient has read a message.

Now, the world’s mobile operators are working on asecond iteration of SMS called RCS (richcommunications services), which keeps its universalitybut enhances the channel with OTT-like features.

Actually, work started first on the new platform in2008, and the project went through various iterationsuntil 2015. It was at this point that Google bought Jibe(an independent company that specialised in buildingRCS services).

Then, in February 2016, the GSMA (the operators’ tradebody) confirmed a formal collaboration between majoroperators and Google to work together on the RCSmessaging standard. Essentially, Google will build asingle Android RCS client based on a universal RCSprofile the operators all agree to. This client will be‘standard’ for Android devices.

Sprint at the startIn November 2016, Google disclosed that more than58 carriers and manufacturers, collectively covering asubscriber base of 4.7 billion people, had committed tothis single, standard implementation of RCS. It alsorevealed that US operator Sprint had launched RCSmessaging to its customers using Android devices

Thus, Sprint subscribers had their standard SMSexperience upgraded via the Android Messenger app.In other words they could use SMS with features suchas group chat, high-res photo sharing (100x larger thanwas possible before), read receipts and more.

On its blog, Google described how RCS improves theuser experience: “For example, though MMS todaysupports group messaging, you can’t do some of thethings people now expect – like naming the group,adding new members, or leaving an existing group.That will change with RCS.”

Default is yoursGoogle added that in 2017, all new Android devicesfrom Sprint will come with Messenger for Android pre-loaded as the default SMS and RCS messagingexperience. And in the coming months, it will launchmore partnerships with operators in other regions.Meanwhile the GSMA says it has a “path to one billionusers” for RCS.

Clearly, the challenge ahead for Google and for theworld’s operators is to meet these objectives. After all,messaging has to be universal. A platform only worksif your friends can use it too.

Next gen SMS?Google it

66 67

MEF FUTURE OF MESSAGING GUIDE RCS IS SMS V2 MEF FUTURE OF MESSAGING GUIDE RCS IS SMS V2

Page 35: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE PREDICTIONS FOR PREDICTIVE MEF FUTURE OF MESSAGING GUIDE PREDICTIONS FOR PREDICTIVE

Swiftkey’s new model for predictive text borrowsfrom the human brain. It’s a step change, saysDr Ben Medlock, CTO of SwiftKey...

People love to text, but they hate to type. That’swhy the history of mobile messaging has seenusers jump at any breakthrough that reduces the

need to tap. They abandoned the numeric keypad forthe QWERTY touchscreen. They embraced emoji andgifs. And they downloaded predictive keyboards.

Perhaps the best known of the downloadablekeyboards is SwiftKey, which launched on Android in2010. Swiftkey’s breakthrough was a three-wordprediction bar above the keys that predicted a user’snext word. Powering the feature was a system basedon words previously entered by the user. This is knownas the N-gram model.

N-gram was a big success. It’s now used on more thana billion devices. But all the while SwiftKey had an evenbetter solution: neural networking. This subfield ofartificial intelligence mimics the human brain in the

way it processes the relationship between words.SwiftKey knew the technology could deliver theultimate goal of reduced typing. But neural networksdevour processing power. They require large servers.For years, mobile processors just weren’t up to thetask. Now they are. So in October 2016, SwiftKeyNeural Alpha went live.

On the flyThe neural networks model captures the relationshipbetween words. Because it understands wordsimilarity, it can compare words on the fly. The bestway to understand how neural networks work is toapply it to actual typing. Say a user is typing ‘this couldtake a…”, the N-gram system might suggest ‘look’because it has previously observed the sequence ’takea look’. The neural model would not make this mistake.Similarly neural networks can suggest words the userhas never typed before. Dr Ben Medlock, CTO of

SwiftKey, says the breakthrough is preparing languageprocessing for a future which moves beyond the phoneand in which machines converse on behalf of humans.

Machine talkHe says: “Even when we launched in 2009, we had theintuition that the creation of language was going to beless about the arrangement of keys and more aboutthe way the software processed the information. Sowhile the N-gram model was a significant idea, wenow need to work towards a future in which deviceswill become more varied and we will need new waysto create language.”

That future is already being built. There is already amarket for wearables and VR, though embryonic. Andthe challenge is on to find ways to ‘talk’ to them.Medlock says: “We need to work out the best way forthese devices to take in a variety of data and signals –

that could be from voice or gesture or touch orwhatever – and make conversations natural. Afterthat, we can begin thinking about a future in whichagents can create language for you.”

Screens foreverHe gives the example of a ‘live’ email ‘out of office’function that responds differently to every messagedepending on the relationship between the sender andthe recipient. “Machines should know you well enoughto do this, like a human PA would. Our neural networklaunch is a step towards this. It’s the logical conclusionof all the tools we’ve been building.”

But don’t assume this means an end to the traditionalkeyboard and screen. “There are limits to voice. If youwant directions or cinema listings, you want to look ata map or menu, not have options read to you,” saysMedlock. “I don’t think screens are going anywhere.”

What’s neural inpredictive text?

68 69

Page 36: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE SPONSORS

70

As market leader in enterprise application software, SAP helps companies of all sizes and industries run better.From back office to boardroom, warehouse to storefront, desktop to mobile device – SAP empowers people andorganisations to work together more efficiently and use business insight more effectively. SAP applications andservices enable approximately 320,000 customers to operate profitably, adapt continuously, and grow sustainably.

www.sap.com

MEF FUTURE OF MESSAGING GUIDE SPONSORS

71

With thanks to ourpartners and theparticipants of MEF’sFuture of MessagingProgramme

CLX Communications AB (publ) connects enterprises to people and things. We combine programmable APIs andcloud computing with our unparalleled Tier 1 Super Network to make it easy for businesses to embed globalcommunications, including voice, SMS and mobile data into their apps, business processes and IoT devices. CLXCommunications AB (publ) is listed on the Nasdaq in Stockholm.

www.clxcommunications.com

OpenMarket, a division of Amdocs, provides mobile engagement solutions for organisations to optimize theiroperations and enhance relationships with customers and employees. Major enterprises choose OpenMarketfor its domain expertise, service flexibility, demonstrated performance, reliability, global scale, and corporatematurity. It provides connectivity to more than 200 countries, enabling businesses to engage with nearly everymobile user around the world.

www.openmarket.com

Page 37: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

MEF FUTURE OF MESSAGING GUIDE CREDITS

72

MEF FUTURE OF MESSAGING GUIDE CREDITS

73

MEF is a global trade body that acts as an impartial and authoritative champion for the mobile ecosystem. Itprovides members with access to networking, collaboration and industry solutions. The MEF’s goal is toaccelerate the growth of a sustainable mobile ecosystem that drives inclusion for all and delivers trustedservices that enrich the lives of consumers. The organisation was established in 2000 and is headquartered inthe UK. It has chapters across Africa, Asia, Europe, Middle East and Latin America.

http://mobileecosystemforum.com

Published by:Mobile Ecosystem Forum

Editors:Tim Green, features editor, Mobile Ecosystem ForumJoanne Lacey, COO, Mobile Ecosystem Forum

Design and layout:155 Creative

Photo credits:Creative commons. Photos and images supplied by the interviewees are published under their responsibility.

MEF’s Future of Messaging Programme is a collaborative industry platform bringing together all stakeholdersin the mobile messaging ecosystem to achieve a common goal to promote and accelerate best practices inorder to limit fraudulent behaviours and identify new opportunities for mobile messaging. Global programmemembers participate in two dedicated work streams Fraud Management and Market Development workingtogether to secure a more transparent, innovative and fraud-free market for the benefit of consumers,enterprises and industry alike; creating a sustainable future of messaging.

http://futureofmessaging.com

Page 38: A2P to OTT, chat bots to smart FUTURE OF PREDICTIVE · INTRODUCTION Mobile messaging now: stats, trends, challenges ... are chat bots - virtual machines that could be the future of

ACCELERATINGYOUR MOBILEOPPORTUNITY

ELERACCAAC TELERAATTING

OPPORYOUR MOBILE

UNITOPPORYOUR MOBILE

YTUNIYOUR MOBILE