Ever since the conception of the Turing test, technologists and designers have been imagining what a truly AI chatbot would look like. The recent developments in machine learning and natural language processing are making this dream a possibility, every day a little bit more than the last one. Not only that: Every day is increasingly clear that a truly intelligent bot will need multiple characteristics and disciplines to reach the final, vowing effect. Such disciplines are not limited only to machine learning and Natural Language Processing. User Experience Design, conversation design, and user research will play a big role.

In this regard, an AI chatbot can be conceptualized as an interface, similar to the Graphical User Interface of phones and desktops, or to the command line of the old MS-DOS era. At the end of the day, it’s another way to interact with a computer. And if you take a look at how the GUI concept was generalized, it was a concerted effort of design and engineering. For example, for the original Macintosh – that popularized the concept of GUI with mouse and desktop – required the artistic work of Susan Kare, as well as software and hardware engineering. As of now, all of the signals point to the fact that the creation of a truly AI chatbot will require the same kind of collective, multi-disciplinary effort.

In our experience, the biggest contribution that a product manager, or designer, can bring to the creation of an AI chatbot is to define its goals crisply. True, in general, this consideration may be true for any project. But in this scenario, the breadth and goals of the chatbot define user expectations, and – together with it – the expected amount of intelligent conversational capabilities that makes it intelligent, as perceived by the user. In other words: It’s very difficult for you to consider Siri truly intelligent because Siri’s goals are so broad that you would expect to be able to tell her anything about any topic. A chatbot that presents itself as the customer support of AirFrance, instead, is expected to demonstrate intelligence only in the domains of ticket booking, new flights, destinations, luggage information. In this regard, the AirFrance chatbot could be perceived as a truly AI chatbot, even though the pure tech behind Siri is more advanced.

So how do you start contributing to this kind of project? Well, there are a million ways, but one of them is to start designing your chatbot with a dedicated tool. To do that, check out Botsociety, we offer a free plan to get you started.


Categories: Ideas