I recently attended a talk by Tim Chang on Messaging is Eating the World, where he discussed how messaging-based interfaces are becoming the new norm in the post-app economy for next-generation mobile services, commerce and transactions.
Messaging has been the fastest-growing online behavior within the social landscape over the past five years, passing social networks, as illustrated below.
Messaging Platform Growth, 2011-2015
Source: Business Insider, Fortune, Mashable, AppAnnie, AdWeek, Quartz, Yahoo Finance, Experian, TechCrunch, Forbes, Tehc in Asia, eMarketer, Compete, Activate analysis
Messaging services are expected to eat both mobile and corporate software
Messaging apps have been disrupting the mobile ecosystem for a while and messaging services are rapidly growing beyond online chat. We actually only use a handful of apps in a day, and the number one is messaging. Today, we can assume that all existing native apps will need to add messaging features.
Messaging is becoming the new platform for everything; including gaming, utilities, entertainment, tracking, etc. After having focused on growth, messaging apps companies are now focusing on building out services for their massive user base.
Chatbots in the spotlight
Chatbots have been in the spotlight lately with Facebook’s new chatbot API for Messenger, Messenger Platform, launched at the last F8 conference. These chatbots will allow businesses to engage with users, and increase the speed and efficacy of interactions.
We are entering a world where we won’t know who (or what!) we are talking to: a dog, a human… or a bot? The real question might be: “Is artificial intelligence (AI) going to eat the world?
But wait… what is the difference between a Chat Interface, a Chatbot and Artificial Intelligence?
Chat app interfaces are targeting human agents. Chatbots (or bots) are computer programs (procedural, scripted, decision trees) that are able to have a conversation with human users. Bots do not necessarily imply artificial intelligence. But more and more of them (AI bots) will use artificial intelligence to communicate. They will rely on Natural Language Processing/Understanding, that often use machine learning and deep learning.
AI bots need training to gain autonomy.
Take the example of Tay, the Twitter AI bot Microsoft launched last March. It was an experiment in conversational understanding: Tay was learning to engage people through casual and playful conversation and the more you would chat with Tay, the smarter it would get. Unfortunately, some users abused Tay’s commenting skills and Tay started responding in inappropriate ways. As a result, Tay was silenced after 24 hours
The ability to define context will be key for the usefulness and adoption of chatbots. The more context the chatbot will be able to define in a short amount of time, the more value it will have. Indeed, if you have to engage in a long conversation with a chatbot for it to understand the context and give you the appropriate answer, you might get tired of it really quickly! The value of the chatbot will decrease with the number of iterations it requires.
Future of chatbots
Regarding user experience, the right amount of push messages will need to be defined in order to increase user acquisition. Ideally, push messages will be sent at the right time and just in the right amount, and then the bot will wait for the user to respond (pull message). “Talent” in messaging will be key: personalized tone, language and emotion will play a big role in user adoption.
We might see the arrival of metabots that will manage a group of bots. Of course, security and privacy will still be major concerns.
AI talent and proprietary data are already the new gold. The GAFAs have all the power and are forcing F500 to keep up, by merging, acquiring and bringing outside innovation in.
In an upcoming article, we will take a look at innovative startups developing bot-based technologies for healthcare, enterprise software, and much more.
Tim Chang is a Managing Director at Mayfield Fund, focusing on venture capital invests in Mobile, Gaming, Digital Media, eCommerce/Marketplaces and Health/Wellness. He has been named twice on the Midas List of top 100 VCs. Tim led Mayfield’s investments in Basis (acquired by Intel), MOAT, 3Drobotics, Massdrop, HealthTap, Fitmob (acquired by Classpass), and Lantern. Before entering the venture capital industry, Tim was a Product Manager at Gateway, where he launched Enterprise products into the Japanese market. Tim received his MBA from the Stanford Graduate School of Business and holds an MSEE and BSEE in Electrical Engineering from the University of Michigan.