conversational ai

20
Conversational AI: A Market Overview

Upload: others

Post on 04-Jan-2022

6 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Conversational AI

Conversational AI:A Market Overview

Page 2: Conversational AI

2Conversational AI: A Market Overview

Executive Summary ........................................................................................ 3

Introduction .................................................................................................... 4

When Messaging, Voice and AI Converge ............................................... 5

The Current State of the Industry: Messaging Apps ........................................ 7

The WeChat Effect: How E-commerce Migrated to Messaging ............... 8

A Less Chatty Chatbot .......................................................................... 10

Why Chatbots? ..................................................................................... 11

Third-Party Bots Accelerate the Shift to Conversational AI .................... 12

The Current State of the Industry: Voice Assistants ....................................... 13

Why It’s Worth Betting On Voice-Activated Assistants Now .................. 14

The Current State of the Industry: AI ............................................................. 15

Microsoft Makes Up for Lost Time ......................................................... 17

Planning For an AI-Driven Bot Evolution ................................................ 18

Conclusion ................................................................................................... 19

Table of Contents

Page 3: Conversational AI

3Conversational AI: A Market Overview

More than 2 billion people use text- and voice-based chat platforms like Facebook Messenger and Amazon Echo. Meanwhile e-commerce and interactive conversational experiences from businesses are beginning to flourish. Chatbots – computer software programs designed to simulate human conversation via text or audio – are emerging as the key mechanism for companies to interact with consumers at scale across these platforms.

Although chatbots often underwhelmed users initially, they will improve rapidly and offer a variety of benefits to business, including capacity for personalization, high engagement via push notifications, productivity via automation, revenue generation and access to consumers on wildly popular apps.

The tech industry is in the midst of an artificial intelligence boom, with titans of

industry like Microsoft, Google and Amazon bringing innovation to the mainstream via various products and cloud services. AI will serve as the engine driving chatbot experiences, and technical progress and lower costs leading to continual improvements in chatbot performance.

Marketers should experiment with bots now. Even if the initial tech employed is fairly simplistic, it’s possible to introduce increasingly advanced AI to a core experience over time. For chatbots to succeed, they must deliver on one of several traits, including efficiency, personality, intelligence, exclusivity and shareability. Simply replicating experiences already found on a company’s website or app is unlikely to retain user interest for long.

To build a strong chatbot team, include the right mix of engineers and creatives, or alternately you may chose to outsource

the work to third-party specialists. Make sure to design experiences that cater to your specific audience and your users’ favorite platforms and features. Chat platform analytics are limited compared to traditional social networks, but you’ll be able to measure and optimize based on metrics like subscribers, revenue, shares, retention, open rates, number of messages exchanged and website visits.

Numerous companies in verticals from banking to recruiting, real estate and entertainment, are already seeing great results with chatbots. Getting an early start on bot experimentation has distinct advantages, including the ability to build a large audience before the market becomes saturated.

Read on to learn more about conversational AI and the exciting opportunities it represents for your business.

Executive Summary

3Conversational AI: A Market Overview

Page 4: Conversational AI

4Conversational AI: A Market Overview

Industry predictions for the rise of artificial intelligence (AI) and conversational interfaces often sound like the stuff of science fiction. Futurists from Ray Kurzweil to Neil deGrasse Tyson describe a singularity on the horizon, where machines become exponentially smarter than humans, ushering in an age of automation that will address such ambitious challenges as aging and mortality. But for C-suite executives in the here and now, there are practical considerations to weigh, and relatively little guidance available to adequately plan for the future. There’s also a growing disconnect between today’s hype-fueled chatbot craze and the underwhelming reality of experiences and tools currently in the market.

Introduction

4Conversational AI: A Market Overview

Page 5: Conversational AI

5Conversational AI: A Market Overview

What’s clear is that many consumers are using messaging platforms to interact with each other and with brands, and that many are also experimenting with intelligent voice applications. Importantly, for those companies wanting to stay ahead of the competition, the technologies to enable hyper-personalized, intelligent and automated interactions are rapidly advancing and becoming increasingly accessible to all.

Our goal in this paper is to give executives an understanding of the market. In a second, forthcoming paper, we will provide an approach for getting the most out of the opportunity that conversational AI represents. While the excitement for AI and conversational user experience (UX) is warranted, it will only become transformative at the business level gradually and over the course of several years.

By exploring the reasons for this slow build, we’ll evaluate what success can look like today as well as what it may look like in five to 10 years. What’s clear is that now is the time to start experimenting, strategizing and building large audiences on chat platforms to future-proof your business and develop new long-term sources of revenue.

When Messaging, Voice and AI Converge

Although marketers often view messaging apps, voice assistant platforms and artificial intelligence as disconnected technologies, they will converge in increasingly important ways. Consumers are already communicating with businesses on their favorite chat platforms like they do with friends and family: using text messaging or voice depending on what’s convenient at the moment.

For example, users should be able to interact with their favorite retailer on both Facebook Messenger and Amazon Echo. And that retail brand should be able to hold a seamless, synchronous conversation with their consumers across both platforms, regardless of whether they’re at home, on the go or in the car.

A chatbot is a computer software program designed to simulate human conversation via text or audio messages. The format initially gained traction in the 1950s and 1960s through the work of visionaries like Alan Turing — whose Turing Test set criteria for convincingly human bot behavior — and Joseph Weizenbaum, whose ELIZA simulated a Rogerian psychotherapist. Although bot technology has evolved significantly over the years, it has maintained core characteristics like an automated backend that retrieves and delivers information in response to inputs from an end user.

“What is a Chatbot?

Page 6: Conversational AI

6Conversational AI: A Market Overview

Whether it’s via voice or messaging interfaces, artificial intelligence is the glue that will power a company’s ability to scale interactions across multiple platforms, leveraging stored information about each user from previous conversations to deliver natural, nuanced and gratifying user experiences. Businesses will eventually manage their presence across all text and voice platforms from a central hub, leveraging additional user data collected from interactions on- and offline.

While this is largely a projection of what could be, it’s already starting to take shape today. To see how, let’s take a closer look at the current messaging, voice and AI landscapes, and how they’re beginning to align.

AI will power your ability to scale interactions across multiple platforms.

Page 7: Conversational AI

7Conversational AI: A Market Overview

Messaging apps have risen to global prominence as a low-cost alternative to text messaging. With Wi-Fi proliferation and higher connection speeds, they have quickly evolved into multimedia platforms built around personal conversations. By late 2015, the social media landscape had undergone a complete sea change, with chat apps overtaking traditional social networks in collective audience size.

The Current State of the Industry: Messaging Apps

7Conversational AI: A Market Overview

Page 8: Conversational AI

8Conversational AI: A Market Overview

As of 2017, over 2.5 billion people are using messaging services, with roughly a dozen major platforms claiming various geographic and demographic strongholds.1 Five of the 10 all-time apps are messaging apps, and 75 percent of smartphone users use at least one chat app.

Even Facebook, the king of traditional social media, has reinvented itself as a chat-first company. Today, there are more than 2 billion active monthly accounts across WhatsApp and Facebook Messenger, compared to 1.94 billion monthly active users on Facebook.2 Meanwhile, 63 percent of Facebook users have increased messaging with businesses over the last year, with more than a billion messages exchanged between consumers and businesses each month on Messenger.3 And Instagram, with its 700 million monthly users, has also introduced a variety of new tools to encourage one-to-one chat between brands and consumers.4

While most major messengers gained popularity with a core differentiating feature, competitors eventually adopted each other’s successful products. Over time, a fairly common set of user tools emerged: texting, video and audio calls, GIFs, stickers and games. Most important, native payments have emerged as effective tools for companies to bring e-commerce to consumers.

The WeChat Effect: How E-commerce Migrated to Messaging

In late 2013, the top apps began building tools for business, with Tencent’s WeChat consistently proving itself the envy of, and inspiration for, other top messengers. With 938 million monthly active users, the app has essentially become a portal to the mobile web in China.5 Chat is its core functionality, but users can also book flights, taxis and appointments, pay bills, buy goods and gifts, monitor air pollution, and perform a host of other activities without ever leaving the platform.

1 “Bots, the Next Frontier,” The Economist, April 2016.

2 “Company Info -- Stats,” Facebook News Room, June 2017.

3 “Fourth Quarter and Full Year 2016 Results Conference Call,” Facebook, Inc., February 2017.

4 Alex Heath, “Instagram’s user base has doubled in the last 2 years to 700 million,” Business Insider, April 2017.

5 Jordan Novet, “China’s WeChat captures almost 30% of the country’s mobile app usage: Meeker report,” CNBC, May 2017.

Customers are adopting new channels, both messaging platforms like Facebook Messenger or Kik as well as voice-based intelligence platforms such as Google Home and Amazon Alexa. Identify where your target customers and users are and meet them on those platforms.

Your Customers Are Moving, Will You?

Page 9: Conversational AI

9Conversational AI: A Market Overview

Today, 50 percent of WeChat users spend 90 minutes per day in the app. More than 300 million credit cards are linked to WeChat Pay, and over 300,000 offline retailers accept payments through the platform.6 At chains like McDonald’s, Chinese customers regularly collect and redeem coupons via the app, and 30 percent of patrons in China have paid for meals via the platform, making it the chain’s top method of payment alongside cash.7

Businesses on WeChat have several key tools at their disposal, including chatbots, microsites and payments. Chatbots typically offer simple decision tree menus, push notifications and the ability to chat with humans for customer service. Microsites — branded mobile sites native to WeChat — can be customized to include e-commerce shops, loyalty programs and a range of other features (more WeChat microsites are launched in China each day than traditional websites). And payments are processed through the app’s proprietary WeChat Pay system.

Western chat apps like Facebook Messenger study the WeChat business playbook carefully, emulating its success to deliver e-commerce and consumer engagement opportunities. Kik, a Canadian messenger with over 300 million registered users, including 40 percent of US teens, went so far as to declare its intention to become “The WeChat of The West.” 8,9

Messenger piloted chatbots in 2015, rolled them out broadly in 2016 with payment integration, and doubled down with a slew of new features at Facebook’s 2017 F8 global developer conference. Messenger’s adoption of chatbots had a ripple effect across the chat ecosystem as rivals mobilized to offer competing products. Today, businesses can launch chatbots on 10 major platforms to engage consumers: Facebook Messenger, WeChat, Skype, LINE, Viber, Kik, Telegram, BBM, Twitter and Slack.

6 Jon Russell, “Messaging app WeChat is becoming a mobile payment giant in China,” TechCrunch, March 2016.

7 “WeChat Pay reached over 300K retail stores,” China Internet Watch, April 2016.

8 Lucas Matney, “Kik already has over 6,000 bots reaching 300 million registered users,” TechCrunch, May 2016.

9 Ted Livingston, “The Race To Become The WeChat of The West,” Medium, November 2014.

China’s most popular messaging app, WeChat is evidence for companies around the world that chat apps can transform mobile commerce. The platform has 938 million monthly active users, including over 300 million consumers and 300,000 retailers using its WeChat Pay e-commerce system. Which messaging platform will become the “WeChat of the West” is still unknown, but CEOs should pay heed to the models that have emerged in China.

What We Can Learn From WeChat

Page 10: Conversational AI

10Conversational AI: A Market Overview

A Less Chatty Chatbot

“We think you should message a business just the way you would message a friend,” announced Mark Zuckerberg at F8 2016. “To order flowers on 1-800-Flowers, you never have to call 1-800-Flowers again.” Those bold assertions inevitably drew tremendous attention, excitement and polarizing opinions from tech journalists, brands and chat app competitors.10

After launching chatbots in 2013, WeChat quickly learned that in the absence of advanced AI, it needed features like persistent menus, microsites and the option to talk to humans to deliver satisfying customer experiences. In April 2016, WeChat product manager Dan Grover wrote a blog post that was highly critical of Messenger for repeating WeChat’s early mistakes:

“Designing the UI for a given task around a purely conversational metaphor makes us surrender the full gamut of choices we’d otherwise have.... [WeChat] added countless features to its APIs — and yet those that actually succeeded in bringing value to users were the ones that peeled back conventions of ‘conversational’ UI. Most instructively, these successes were borne out of watching how users and brands actually used the app and are seeking to optimize those cases.”11

Messenger appears to have taken the critique to heart; Facebook poached Grover from WeChat soon after. By F8 in April 2017, it pushed bots in a decidedly less chatty direction, encouraging developers to use WeChat-like persistent menus for navigation, and web browsers that open brand sites within the chat thread.

Messenger also rolled out chat extensions designed for companies to facilitate consumer-to-consumer interactions and group conversations around their product or service. With this model, launch partners like Spotify, Kayak and Open Table offered in-chat apps that enable friends to find and share music, plan trips or book restaurant reservations. The approach also mirrors Apple’s late-2016 launch of “iMessage apps,” with experiences from companies like Jet and Fandango facilitating in-chat product and movie ticket purchases between friends.

10 Jessica Guynn, “Zuckerberg’s Facebook Messenger launches ‘chat bots’ platform,” USA Today, April 2016.

11 Dan Grover, “Bots won’t replace apps. Better apps will replace apps.,” DanGrover.com, April 2016.

Until businesses can access affordable advanced AI, chat apps are enhancing conversational experiences with user-friendly features like suggested response buttons and mobile sites that open within the chat thread.

Bots Are Becoming More Like Apps… For Now

Page 11: Conversational AI

11Conversational AI: A Market Overview

Similarly, platforms like Slack and Twitch have popular bots geared toward productivity and efficiency that minimize actual chat to focus on completing specific tasks within team or community conversations. For example, Slack’s users that enable the Zoom bot can simply type “/zoom” in a conversation to initiate an instant Zoom video conference or call.

Without inexpensive advanced AI widely available to businesses, expect bots to become increasingly app-like over the next 18 months, and to increasingly play a support role by enhancing conversation between friends. But, as we’ll see below, truly conversational bot experiences are also beginning to flourish on voice-activated platforms and will eventually return in new-and-improved form to text-based chat platforms.

Why Chatbots?

Studies indicate corporate adoption of chatbots currently lags behind consumer openness to the technology, and that businesses would be wise to pick up the pace. According to a survey by Retale, “59 percent of US Millennials and 60 percent of US Gen X-ers have used chatbots on a messaging app” and chatbots have higher long-term retention rates than traditional apps.12 Yet Forrester reports that as of late 2016, only 4 percent of businesses launched a bot.13

The best bots in the market today accomplish seven key business goals:

1. One-to-one Conversation at Scale. Leveraging automation, businesses can carry on thousands or millions of simultaneous ongoing chats using the app’s native UX.

2. Personalization. With a rich trove of data, companies can tailor messaging at the individual level based on interests, past behavior and responses within the bot conversation.

3. High Engagement via Push Notifications. Once positive experiences have been delivered, businesses can leverage push messages that go directly into user inboxes. Marketers, in turn, can rely on high open rates and an effective ongoing mechanism for re-engagement.

12 Alison McCarthy, “If Anyone Is Going to Buy Directly from a Chatbot, It’s Millennials,” eMarketer, February 2017.

13 Mark Sullivan, “After Lots of Talk, Microsofts Bots Show Signs of Life,” Fast Company, May 2016.

1

2

3

Chat apps have attracted billions of users, and bots are the mechanism that enable organizations to deliver highly personalized interactions at scale. Improved personalization delivers a more relevant user experience, and the high levels of automation used by bots enable cost-effective delivery at scale. As a result businesses can engage with customers and achieve open rates that typically outperform email and social networks.

The Case for Bots

Page 12: Conversational AI

12Conversational AI: A Market Overview

4. A Cure for the Brand App Blues. ComScore reports that smartphone users spend 80 percent of their app time in only three apps.14 With multiple messengers topping app charts globally, having a piece of digital real estate on those platforms is an appealing alternative to building owned brand apps.

5. Revenue. AI-driven chatbots can lead consumers through the entire sales funnel from awareness to purchase.

6. Efficiency and Productivity. Bots have the capacity to enhance conversations between users by surfacing helpful information, or completing repeated tasks like scheduling.

7. Ambient Chat. Major brands from GE to BMW are integrating chat functionality into internet-connected devices, with bot technology powering those consumer interactions.

Third-Party Bots Accelerate the Shift to Conversational AI

Chat platforms, such as Slack, have chosen to rely heavily on third parties to provide bot-based extensions to their core platforms. There’s a long list of startups offering their take on how businesses should build and syndicate bot experiences across multiple chat apps while collecting helpful performance data. Platforms like Chatfuel, Sequel and Motion.ai are geared toward the lower-cost, self-service market. Others like Msg.ai and Imperson are geared toward pushing the envelope on AI, natural language processing and personalization.

The tech giants have also stepped into the fold. Microsoft launched a bot framework that deploys experiences to most major chat apps, including ones competitive to its own chat properties. Google has taken a similar approach through the acquisition of bot firm Api.ai, which supports Facebook Messenger, Slack and Google Assistant. IBM’s Watson Conversation powers chat on Messenger, Slack and Twilio, with additional emphasis on web and internet of things integrations. Amazon Lex can also be used to make bots for Facebook Messenger and Alexa, with Slack and Twilio coming soon.

Recently, many third-party platforms have also introduced tools for creating custom behaviors, as well as enabling, building and launching bots on voice-activated platforms. Google Home has launched actions while Amazon Alexa and Microsoft Cortana both support skills.

14 Victor Luckerson, “This is how many apps you’re really using on your smartphone,” Fortune, September 2015.

4

5

6

7

Traditional social networks provide businesses with self-service official pages, but chatbots are by nature more complex and dynamic than pages. For that reason, chat apps hosting bots still require businesses to build their own conversational agent, or to work with third-party companies specializing in chatbots. While building a chatbot is still complex, the technologies and tools for building bots are improving rapidly.

Chatbots vs. Official Pages

Page 13: Conversational AI

13Conversational AI: A Market Overview

While the big selling point of messaging apps is immediate access to a pool of billions of users, the voice assistant market requires comparative patience. Amazon Echo, the far-and-away market leader, has only sold roughly 11 million units to date according to Morgan Stanley.15 And, eMarketer projects 35.6 million US consumers will use a voice-activated device at least once a month in 2017.16

15 Ángel González, “Amazon has sold more than 11 million Echo devices, Morgan Stanley says,” The Seattle TImes, January 2017.

16 “Alexa, Say What?! Voice-Enabled Speaker Usage to Grow Nearly 130% This Year,” eMarketer.com, May 2017.

The Current State of the Industry: Voice Assistants

13Conversational AI: A Market Overview

Page 14: Conversational AI

14Conversational AI: A Market Overview

On a technical level, the divide between text- and voice-based chatbots is modest. They both rely on artificial intelligence or simple decision tree structures to carry on conversation with users. One can point to multiple popular text-based bots that have been repurposed for voice.

The notable difference between voice and text, from the consumer standpoint, is primarily financial. Billions of people quickly flocked to chat apps because they’re low-cost alternatives to SMS, but growth for hardware at the $180 and $50 price point (Amazon’s Echo and Echo Dot, respectively) cannot achieve similar ubiquity so rapidly.

That picture becomes a bit sunnier if you include voice assistants accessed via mobile. Ninety-eight percent of iPhone owners and 96 percent of Android owners have tried Siri and Google Assistant, respectively. But until Apple and Google invite businesses to build voice-activated bots on those mobile platforms, opportunity is scarce. Amazon recently made Alexa available on its mobile app, but it’s unclear if users are taking advantage of the feature.

Why It’s Worth Betting On Voice-Activated Assistants Now

While voice-activated assistants represent a nascent market, consumers’ enthusiasm is notably strong. Through a combination of word of mouth and extensive TV ad campaigns, Amazon Echo, and to a lesser extent Google Home, became mainstream hits during the 2016 holiday season. A KPCB study shows that consumers are particularly drawn to the technology’s speed compared to text, and its hands-free and eyes-free nature.17

Early adopters of Echo gravitate most to activities like listening to music, news, and setting alarms and timers. But 45 percent of users report adding items to shopping lists on the device, and 32 percent have used it to purchase products on Amazon Prime.18

Numerous small and large companies are noting these encouraging signs, and building third-party “skills” (Amazon’s word for third-party Echo bots). Notable examples include the Starbucks “Reorder” skill, which learns your favorite order so you can request “the usual” any time to pick it up at a nearby location.

17 Niall McCarthy, “The Meteoric Rise of Mobile Voice Assistance,” Statista, November 2016.

18 Felix Richter, “What the Amazon Echo Is Actually Used For,” Statista, October, 2016.

It’s important to remember that voice and text will be appropriate in different contexts. But together, they provide a rich set of options for crafting the right interactions with your customers. Aim to build a seamless experience where customers can chat with your bot however and wherever they want to. New bot-building platforms are emerging that help organizations design, deploy and manage experiences for both formats, such as Microsoft’s Bot Framework, from within a single tool.

Unify Voice and Text

Page 15: Conversational AI

15Conversational AI: A Market Overview

Recent advances in processing technology have led to a boom in artificial intelligence. Tech titans like Apple, Amazon, Baidu, Facebook, Google, IBM and Microsoft are engaged in an arms race. Each is allocating tremendous resources to drive AI progress, and to turn ideas that were once derided as fantasy into reality. Most are acquiring startups with specialized tech and expertise, committing tremendous capital to make technical breakthroughs, and at times rolling out those innovations to the public through consumer products and cloud services.

The Current State of the Industry: AI

15Conversational AI: A Market Overview

Page 16: Conversational AI

16Conversational AI: A Market Overview

Banks like Capital One can recite account balances and tally expenditures, and on the future-gazing front, automakers like BMW and Hyundai offer Knight Rider-esque integrations where you direct Alexa to start your car.

Tesla’s self-driving cars, Amazon’s Alexa and Apple’s Siri are all examples of advanced AI in action, built over long periods of time and continually optimized through machine learning. And while users can easily expose flaws in all of these experiences, at their best they exhibit a magical quality we often associate with transformative technologies.

If you’re wondering when smaller businesses will affordably access these technologies in a way that’s transformative for their own businesses, it’s already happening. A closer look at one of the tech giants reveals where the best opportunities lie today, and what to expect in the years ahead.

Machine Learning. The process of training computers to perform tasks with a combination of data analysis and probability rather than following a strict rules-based approach. This allows the computer to achieve a certain level of independent “thought,” which is typically applied in the form of narrow AI (meaning specific, limited functions like computer vision that enable technology like self-driving cars). However, we are still far off from the general AI of the C-3PO and Terminator variety.

Deep Learning. Also referred to as neural networks, deep learning is a branch of machine learning that emulates the interconnection of neurons in the brain. This process creates a complex web of discrete machine learning layers that all connect to, and impact, each other. As an example, different layers of a neural network may be applied to analyzing various components of one traffic light.

Natural Language Processing (NLP). In recent years, deep learning techniques have been increasingly applied to this core field of computational linguistics that underpins technologies like Siri, Amazon Alexa and Google Home, as well as text-based bots. NLP includes two subset fields — natural language understanding and natural language generation — both of which are seeing continual improvements in language translation, modeling and word parsing.

AI Glossary

Microsoft reports that roughly 25 percent of users have said “I love you” to its Xiaoice chatbot.

Page 17: Conversational AI

17Conversational AI: A Market Overview

Microsoft Makes Up for Lost Time

After getting a late start on the shift to mobile, Microsoft shows no sign of making the same mistake with artificial intelligence. Its blossoming multi-pronged AI empire is built around an Artificial Intelligence and Research division that’s working to “democratize artificial intelligence” and build “the world’s most powerful AI supercomputer… making it available to anyone, via the cloud.”

Today, Microsoft’s clients can access AI in various forms through a suite of tools like Cognitive Services, which includes machine learning, language, speech, vision and search processing products. Uber, for example, uses facial recognition via Microsoft’s Face API to ensure each driver using its app matches their file on record.

“Integrating this API into our platform took about three weeks even though it was still a preview product at the time,” says Uber’s product manager Dima Kovalev. “Because the API has matured since then, it would probably take even less time today. It saved us months of development work in building face detection capability into our platform. We were able to implement our vision at warp speed... That left us more time to spend optimizing the user experience.”19

Microsoft’s Bot Framework helps businesses build and deploy chatbots to its Cortana virtual assistant, messaging properties like Skype and GroupMe, and competitor platforms like Facebook Messenger and Kik. Clients like StubHub use the Bot Framework to launch both a Skype chatbot and a voice-activated skill on Cortana. Those experiences also tap Microsoft’s proprietary natural language processing (NLP) AI to enhance conversational ability and pull data from StubHub’s servers. The end product puts StubHub in front of a new audience that can ask questions like “Cortana, ask StubHub to find me something fun to do tonight.”

Microsoft also builds its own commercial chatbots, and is responsible for the market’s biggest success: Xiaoice. Built for Chinese users, the bot has over 40 million registered users across multiple social platforms like WeChat. Xiaoice simulates a 17-year-old female to deliver digital companionship and uses advanced AI to power its conversational skills.

19 “Uber boosts platform security with the Face API, part of Microsoft Cognitive Services,” Microsoft Enterprise.

Microsoft is making significant investments, including acquisitions, to build out both artificial intelligence and conversational capabilities. Its Bot Framework is a credible platform that promises to enable developers to work across a wide range of messaging and voice platforms, as well as embed bot functionality in a custom application.

Microsoft: Not Over, or Out

Page 18: Conversational AI

18Conversational AI: A Market Overview

The bot can track users’ emotional states and recall personal information individuals share in conversation, such as if they got a new job or had a bad day. It even offers a breakup therapy course for users grappling with romantic woes. Microsoft reports that roughly 25 percent of users have said “I love you” to the bot, with people texting with it an average of 60 times per month on apps like WeChat.20 In fact, Xiaoice became so popular that a TV network had her start reading the weather report on-air.

Planning For an AI-Driven Bot Evolution

If voice and text-based chatbots were automobiles, then AI would be the engine powering those conversational AI experiences. The car may be basic to start, but over time can be souped up with modifications and customizations to transform it into a high-performance vehicle. Furthermore, there’s plenty of value and mileage to get out of the bot at all points on the journey, and learnings along the way inform smarter decisions as you increase capability.

Taking Microsoft’s clients as an example, every technical advance in the Bot Framework and Cognitive Services means potential improvements for their existing bots built on those platforms. In fact, Apple takes this basic approach with Siri, and in an interview with Mashable described the process of giving Siri a “brain transplant” — replacing the simple rules-based system used at launch in 2011 with a new machine learning and voice recognition infrastructure.

20 Nicky Cappella, “Microsoft’s Chinese social experiment: the largest Turing test in history,” The Stack, February 2016.

The sophistication of your bot is largely determined by the sophistication of the artificial intelligence you employ on the backend. Tech companies like Microsoft, Google and Amazon are working hard to democratize advanced AI so that businesses can tap into their capabilities to introduce increasingly complex experiences.

The AI-Chatbot Connection

Page 19: Conversational AI

19Conversational AI: A Market Overview

Conclusion

19Conversational AI: A Market Overview

Despite some early setbacks during initial bot platform launches, conversational AI opportunities on chat platforms can’t be ignored. Messaging companies and third-party tool providers are doubling down to make rapid improvements. And, given the massive audiences and natural and intuitive nature of the interface, experiences like bots, whether text- or voice-based, are here to stay.

The AI and automation boom is making it possible to carry on unique conversations with customers at scale, delivering increasingly satisfying experiences that drive engagement and loyalty with customers across a wide variety of business functions. There is a growing suite of tools available to support the development of conversational interactions to enrich bots and imbue them

with depth. As software entrepreneurs, product leaders and marketers, our job is to start simple, test and repeat.

The industry is still in an experimental stage, but one that’s evolving quickly. Bots will become more app-like with menu systems to provide some predictability for users until affordable advanced AI becomes pervasive. Over the long term, conversational experiences have the power to transform your business and relationships with your customers.

The sooner you start building a nimble team armed with the right skills and strategy, the bigger the advantage your company will have in the market. Read more about conversational AI in the 9 Principles of Conversational AI.

Page 20: Conversational AI

20Conversational AI: A Market Overview

[email protected] // georgianpartners.com

About Georgian Partners

Georgian Partners is a thesis-driven growth equity firm that invests in SaaS-based business software companies. We look for companies that use foundational technology trends such as applied artificial intelligence, conversational AI and security first to dominate their markets.

Founded by successful entrepreneurs and technology executives, at Georgian Partners we leverage our deep software expertise to directly impact the success of our portfolio companies. That expertise spans areas as diverse as machine learning, analytics, deep learning, cryptography, linguistics, natural language processing, differential privacy, software engineering, information security and cloud computing.