Growth Spotting: Conversational Interfaces
Conversational interface, the ability for users to interact with computers in natural human language, was one of the most hyped technologies last year. Articles circulated online describing how bots, in the future, will help us with everything from booking meetings, sending timely weather notifications, curating news, tracking packages, to ordering flowers for our loved ones.
In the scenario that the technology lives up to these expectations, it will be a transformative change in how we shop, consume information and interact with devices around us. In the long run, this shift would impact products and services from B2C and B2B companies. This is the reason why we’ve shared this article so that you can get a better understanding of where the technology is today and the direction in which the technology is headed. This article will be particularly useful to product owners, entrepreneurs and executives shaping the direction of future consumer products and services.
One of the aspects that separate conversational interfaces from other hot tech topics, such as blockchain and big data, is the fact that the concept of communicating with a computer in natural language has been around since the 50’s. Therefore, it’s by no means a new concept. However, it’s a concept that is made possible and commercially intriguing right now because of the convergence of three main technology trends:
- First, advances in deep learning empowered by large data sets have drastically improved Natural Language Processing (NLP). This allows computers to understand human language nuances with uncanny accuracy.
- Second, messaging apps have surpassed social media in terms of monthly active users (see the graph below). As users are spending an increasing amount of time on messaging apps, it becomes more attractive for organizations to be available via messaging platforms.
- Third, with the rise of mobile, always-on consumer, companies are increasingly expected to deliver fast, reliable customer service across different channels. Messaging bots are a new way for companies to meet these rising expectations.
As seen on the graph, monthly active users on messaging apps surpassed users on social networking apps in mid-2015
In this article, we will take a look at the maturity of the technology by analyzing the strategic actions and positions of the main technology giants (Facebook, Amazon, Apple & Google). Further, we will also explore the traction by WeChat, a messaging platform created by a Chinese tech giant Tencent. Before anyone else, WeChat was able to combine a communication platform with direct in-app transaction (order food, movie tickets, clothes, etc.). Many tech bloggers argue that the big Western tech players are all fighting to become the WeChat of the West or as WeChat is also referred to: “The one app to rule them all”.
As data is the fuel to developing a superior digital assistant, it is no surprise that each of the mentioned companies is fighting to collect as much information as possible. In fact, Google and Facebook collect so much data on their users that people sometimes joke that the companies know their users better than the users know themselves. In terms of the future, it is unclear how the privacy regulations will develop and impact data collection and usage, but both Facebook and Google have been taken to EU court regarding their massive collection of user data.
During the F8 conference last April, Facebook announced that developers now were able to build chatbots to the 900 million active users on the Messenger platform. Developers took interest and after six months, 11,000+ bots were developed and launched on the Messenger platform. However, the response by consumers was underwhelming as many reported about buggy experiences. This partly reflects the low entry barriers to creating a Facebook bot, but also that the technology and guiding principles for bots have still not matured.
Today, almost a year after the conference, we have yet to identify a bot which has been a category winner. Here are some examples of different third party apps that nonetheless have gained some traction on Messenger:
- VisaBot (http://visabot.co/) – the world’s first immigration bot. This bot will help you assess your chances for particular US visas. Further, the app allows you to connect with an immigration lawyer. The company was just accepted into 500 Start-ups.
- KLM bot (https://messenger.klm.com/) – allows you to receive travel information through a bot. You can receive notifications, check-in and print a digital boarding pass. The app serves as an easy accessible customer service application.
- Fynd bot (https://fynd.me/) – a personal shopping- and customer service bot. With Fynd, you have a shopping agent that finds products based on your needs. Further, Fynd is also able to handle returns and provide you with delivery dates.
Facebook is also working on its own digital assistant called M. The purpose of M is to become a universal bot, which is able to handle all kinds of request (similar to Siri). M has been limited to a small selected user group, but according to Messenger VP David Marcus, M will be made available to more users in early 2017. When M is official to the public, it will be interesting to see if Facebook is able to take advantage of its unique combination of user data: Not only do Facebook have insights into who we are and the content we like to consume, Facebook also have a similar knowledge graph of our closest friends and connections.
Screenshots from the KML Facebook chatbot
Amazon’s strategic position differs from that of Facebook, Apple, and Google, in that they don’t control any messaging platforms or smartphone distribution (Amazon did try to enter the smartphone market with Amazon Fire, but the effort failed miserably).
Given this position, Amazon released a consumer hardware device called Echo back in 2014. The Echo is a speaker accompanied with a voice service called Alexa, which can answer general questions like Google or place orders directly through the Amazon store. The release of a voice assistant hardware device was made years before competitors. This shows how badly Amazon wanted to get a device into the lives of consumers and its willingness to make big bets on unproven concepts.
Unlike a smartphone, the Echo does not include a screen interface. However, the Echo has provided Amazon the chance to connect with consumers outside the virtual world. Recently, Amazon released the Echo Dot, an inexpensive alternative to the Echo, at the cost of $49. According to Morgan Stanley, Amazon sold more than 11 million Echo units per December 1st 2016. Sources also indicate that owners of an Echo on average increase their spending on Amazon. This could be the reason for why Amazon are pushing its hardware at such affordable prices.
Another interesting fact is that Amazon is opening Alexa up to third party devices and apps. During this year’s CES several devices came integrated with the Alexa assistant. Some examples include smart refrigerators and cars. Further, Alexa also allows companies to integrate their services directly into the assistant. The additional services are known as Alexa Skills and require users to activate each particular skill before it can be used. An example of an Alexa skill is to book a ride via Uber. By June 2016, Amazon announced that more than 1,000 third party services had been integrated. Partners include companies such as Domino’s, Fitbit, and Kayak.
The entire Amazon Echo product family
In May 2016, Google announced the release of a voice assistant. Surprisingly, Google did not make the assistant available to all android systems. Instead, Google limited the assistant to its premium Pixel smartphone and the Google Home hardware products that compete directly with the Amazon Echo. Google’s decision to keep the assistant restricted to its own products, forced android smartphone-maker Samsung to figure out how to integrate a voice assistant into its future phones. Samsung felt threatened by Google’s decision and acquired the company Viv, an AI-assistant developed by the creators behind Siri, shortly after Google’s announcement.
Initially, the Google Assistant didn’t allow for third parties to integrate their services (Uber, Dominoes, etc.). This changed in December 2016, two months after the announcement of the assistant, as Google made the strategic decision to open up to third-party services. The new third party commands enabled by the Google Assistant are called Google Actions. Some analysts argue that the war to create the best voice assistant is critical to Google, as the voice assistant could become a substitute for traditional web search. This prediction remains to be evaluated.
Examples of third-party services that have been integrated with Google Assistant
Traditional to Apple, the company has not diversified its efforts within conversational interfaces. In fact, Apple’s sole focus has been on its voice assistant Siri which is now accessible for all Apple products – Macbooks, iPads, and iPhones.
That said, Apple surprised many when they announced that Siri will be opened up to third-party services last year. More specifically, Apple will allow developers to integrate the following six services into Siri: Audio/video calling, messaging, sending and receiving payments, searching photos, starting workouts, and booking rides.
Personally, I’ve experienced a significant improvement of Siri – although only in English. It seems that Apple is keenly aware of their deficiencies and they have recently made acquisitions such as Tiri to improve its technology. At the moment, it is still unclear what data Apple will take advantage of to build their assistant. As Apple, in the past, has been known for a less aggressive data privacy approach compared to the approach of Google and Facebook.
Siri’s integration of additional services with iOS 10
While Microsoft entered the mobile revolution too late to catch up with Android and iOS, it doesn’t appear that Microsoft is holding back with conversational interfaces. Cortana, the name of Microsoft’s voice assistant inside Windows 10, has already been licensed into Nissan’s and BMW’s coming car models. Similar to Google, Microsoft announced that their voice assistant opens up to third party services in February.
One of the most intriguing strategic advantages of Microsoft is that most businesses use their productivity tools (email, word, spreadsheets, etc.) and that they acquired LinkedIn, the world’s largest professional social network. While its competitors seem to aim for the consumer segment, Microsoft has a unique opportunity to leverage its assets and deliver the best assistant for businesses. An example use case could be that Cortana helps to seamlessly schedule a meeting between you and a new LinkedIn business connection. On the day of the meeting, Cortana could help you find the appropriate material and give you a quick briefing of the past conversation.
Nissan displayed a demo of Microsoft Cortana at this year’s CES
Tencent is the parent company behind WeChat and is known for digital services in e-commerce, multiplayer online games, and social networks. If you’re not from China or Southeast Asia, you might not have heard about WeChat. That said, WeChat has roughly the equivalent monthly active users as Facebook on mobile devices. Moreover, WeChat leads the trend of expanding a messaging platform to integrate other services (as mentioned earlier in the article). In fact, since 2013 WeChat users have been interacting with bots to customize Nike running shoes, pay at Starbucks or book movie tickets. By 2015, the number of third-party apps on WeChat was estimated at 10 million.
Initially, WeChat began by opening the platform to selected partners. However, WeChat recently decided to allow anyone to develop bots on its platform. This shift has lead to some difficulties; A TechCrunch article reported about the emergence of fake bots, disguising as real users, whose sole purpose is to make new friends and gain a large following. The article also revealed that these bots are sold on a black market. The estimated value for a bot with 5000 friends was between $14 and $140 depending on the quality of the account.
Some of the third-party services that are integrated into WeChat
The largest tech companies are all betting on conversational interfaces, but they are approaching the technology strategically differently. In particular, they differ when it comes to allowing third-party hardware and software products access to their general voice assistant. While Apple, Facebook, and Google are restricting the assistant to their own hardware products, Microsoft and Amazon are allowing anyone from car to fridge manufacturers to integrate their voice assistant.
That said, there are some also similarities: In terms of integrating third party services (Domino’s, Uber, etc.) to their assistants, every company, except Facebook, has decided to allow developers to integrate services with their assistant. In terms of integrations (in the Western World), Amazon is by far the leader at the moment, as they were two years ahead of the other companies. Apple differentiates itself slightly, as it is the only one company that has restricted services that integrate to its assistant to six distinct categories. The decision to limit new services might make it easier for Apple to properly vet new services, but will also come at the cost of limited functionality in the beginning.
Below is a table illustrating the strategic position of each company:
As a final comment, I want to emphasize that this article gives you an overview of the status of conversational interfaces as of February 2017. Expect a lot of development to happen over the coming 12 months as the technology adoption is still in its infancy.
If you have any questions regarding how your company should strategically approach the possibilities of conversational interfaces, you’re more than welcome to reach out.