People have got chatbots all wrong, but that’s not surprising. After all, the ones we’re traditionally confronted with tend to be frustrating tools, delivering confusing responses and displaying a woeful lack of understanding (Remember Microsoft’s Clippy?). That has largely been to do with the compulsion to dress chatbots up with a human voice, to try and mimic human interaction when the technology isn’t yet able to deliver on that promise. As a consequence, we’ve come to believe that chatbots waste our time — but implemented properly, they do precisely the opposite. They can relieve us of onerous or repetitive tasks that we perform every day within browsers and apps; a simple text command can quickly delegate a task such as calling a taxi, booking a flight or tracking a package. Instead of us overseeing each of these things, gazing at screens, waiting for data to be processed and pages to load, we simply let the chatbot get on with it and report back to us when it’s done. In the meantime, we can get on with doing something much more rewarding.
A few years ago, we relished taking personal control of things like, say, booking hotels or renting cars, bypassing travel agents and searching long and hard for the best deals in the belief that it saved us money. But life admin is rarely fun, and surveys continually demonstrate that we’re often overwhelmed with the choices available to us online. The good news is that many of these procedures can be easily automated. Relatively simple code, masquerading as a chatbot, can carry out these tasks, learn from our behaviour, understand what we want and liberate us from drudgery. Disorientating websites and needlessly complex apps can be replaced with simple, functional, two-way exchanges. Your flight can be booked, your dinner reservation can be changed, your bank balance can be checked — each within a single conversational thread instead of across multiple apps, websites, emails and notifications.
This is going to prompt two fairly big changes; firstly a migration away from traditional websites and apps, but also a reduction of the time we spend interacting with services, and less expectation of instant results. Messaging is an asynchronous medium, so we won’t necessarily be expecting a chatbot to deal with everything immediately. It could be in a few minutes, or even an hour or two, as long as we know it’ll get done. Just like any kind of delegation, in fact.
At this early stage of chatbot development, the last thing we should be doing is making bad attempts to showcase what they’ll be able to do in the future. There’s no doubt that within a few years they will develop into fully-fledged personal assistants, wholly dedicated to making our lives easier. But right now, the focus needs to be on the problems we’re having using online services and creating chatbots to solve those problems. The more good-use cases that get made, the more people will want to use them, the more businesses will want to create competing chatbots that work better and deliver more.
So let’s shrug off our chatbot prejudice. Let’s not hold any of their past failings against them, and let’s not allow our fascination with AI — so neatly depicted in films like Her and Ex Machina — to obscure the incredible potential they have, even in the very short term, to simplify and streamline our lives, helping us interact with services and businesses in a much simpler, more efficient, less gruelling way.