Why we are still a long way from a ‘killer’ chatbot

It all started with a conversational interface, and then bots just took over the debate. Even in this great piece, while Dan Grover mentioned the ubiquity of chat, he also highlighted the fact that we should recognize the limitations of pure text-based input method.

Looking at the history of bots, I struggle to see bots overtaking the streamlined GUI experience even now — here’s one recent subpar experience out of many. There’s a good reason why we accepted the web and made it part of our everyday lives.

It’s very naive and sometimes bordering funny to think that bots will be panacea to everything we do in life and will take over humanity one day.

The reality is we’re nowhere near that and perhaps never will be.

Let’s face it  —  bots are simply going to augment our conversational experience, not replace it, and that’s the context within which we should be developing bots. Period.

Chatbots by themselves can’t solve a huge problem of “replacing” humans due to limitations of technology and text based input method. However, when combined with great graphical and conversation design they can be powerful weapons and offer superior experience for many day to day tasks.

At Bottr, we believe a healthy mix of GUI and CUI (conversational user interface) will survive ,  offering a neat and unified experience to access various kinds of information and services.

There are various reasons why many bot developers are getting it all wrong.

First and foremost, the tech is not advanced enough. “When all you have is a hammer everything looks like a nail” could not be more apt than it is in the context of bots and AI. With the hype around bots, everyone is running around like headless chickens to build bots that will solve every existing problem in the world. I predict we will soon see a bot that solves world hunger!

What we fail to realize is that underlying technologies like A.I./NLP (natural language processing) are not advanced enough to give us that natural conversational experience for everything, and we often end up with setting wrong expectations and delivering half-baked experiences.

The real value will come from understanding the limitations of the technology and designing meaningful experiences that can handle the scope of a well-defined problem. This requires a strong command of user behavior and interaction design.

Another major issue with bots is that existing messaging and bot platforms don’t offer an end-to-end solution. Whether it’s talking to a shopping bot or finding information about something, after basic IVR (interactive voice response) style Q&A, customers are often sent to a third-party shopping site or someone’s webpage to get things done. That defeats the purpose of having a bot in the first place: not having to leave the context of what we were doing.

What would be ideal is to focus on experiences that can be best addressed within the remit of a bot to allow for instant gratification and have a seamless handover to a human within the same window if possible when a fail event is detected.

Similar to successful mobile apps such as Uber and Whatsapp, we will see breakthrough business models that are native to chatbot experience. I haven’t come across a single developer who has really thought about how building a weather bot or news bot would offer many times better of an experience to the user and could lead to a stable revenue stream in future.

It’s best at this stage to look at existing problems that have the potential to be solved mainly through bots.

I’m very upbeat about future of chatbots in general and the scope they offer. However, they need to be developed and managed carefully or we will just see a proliferation of meaningless bots that will make users wary of trying them out and gradually lead to fading of developers’ interest in this amazing tool for the future.