Chatbots in 2017

Let’s take a quick look what we can expect from chatbots in 2017.

2016 was a busy year at CartSkill as chatbots made their first mark. With Facebook opening their Messenger platform for developers in April 2016, the bot-rush started. Other large players jumped on the bot wagon as well. Namely Google with its purchase of api.ai and release of Assistant and Home, Apple’s iMessage apps and Amazon with its Alexa skills store for Echo devices.

Here is what we can expect from chatbots in 2017:

1. Chatbots will seek to provide more value for users. In 2016 we saw a myriad of bots that were not very helpful for users. Like the first year of the apps in the iPhone AppStore, the bots were mostly very rudimentary in their usefulness. Expect developers to start focusing more on actual value that bots can offer for users.
2. The big tech companies will march forward with bots but from different angles. Facebook will continue popularizing its chatbots and hoping that brands will warm up more to the usefulness of bots. Amazon will continue offering Echo as the voice controlled bot for anything and may actually come out with headphones to further widen its use. Google will focus on Assistant and the Home product to help gain more widespread adoption for its services. For Apple the AirPods will be key, allowing wireless always on (in-ear) connectivity to Siri. This will, hopefully, teach people how to interact hands-free with Apple devices and services. As you can see it will be a year of learning for users.
3. Various distribution and discovery methods will emerge. More emphasis will be put on ways how people can discover great bots. Existing bot discovery networks will include more ranking and reviews for bots. Also sharing of bots will get easier (e.g. inside group chat) but viral-only distribution strategies clearly do not cut it for developers.
4. Monetization for chatbots needs to be figured out. As a developer there currently is not clear cut monetization model for B2C bots. Unlike apps which have a revenue share model, bots currently lack such a model. This does not provide much incentive to develop chatbots. Expect to see ad networks emerging for bots and more curated bot discovery networks powered by advertising.
5. Intent detection will be key. Current Artificial Intelligence with LSTMs and neural networks has gotten us to 90% understanding of what people are saying and what they want but the 10% we are missing in understanding accuracy is key. Companies will be pushing for solutions in this area with a variety of useful models (including yours truly at CartSkill :).
6. Certain B2B business models will take off. Travel booking; appointment booking in calendars; customer support and certain forms of e-commerce product discovery will find their way into brands’ existing apps and services.
7. In short – experimentation with bots will continue as the space is just getting started and first sustainable models will emerge during 2017.