This post is a follow-on to our recently released FinLab Snapshot, a report in which we identified industry insights, trends and analyses based on FinLab’s 356 applicants in 2016. In this post, we’ll share more about something we called out as one of the “trends on the horizon” we observed: the rise of fintech chatbots.
This year, we saw an influx of chatbot services in our application pool. While the most well known of these products in fintech, Digit, was in our first class of FinLab, this was the year in which we saw a significant volume of other fintech products that followed the same user interface. Roughly 10% of our applicants rely on messaging – SMS, Facebook Messenger, Slack, or custom built messaging apps – as their primary form factor.
The chatbot interface has one huge advantage: simplicity. There’s often no app to install—just send a text message to a number. Even if there is a download, it’s typically very bare bones. There’s a bonus here for low to moderate income consumers that might go unnoticed by those of us with newer phones: reduced storage footprint. Low-end iPhones and Android devices often have just 16GB of storage. With the OS taking up nearly half of that, there’s precious little space left over for apps. Having a slim installation can dramatically lower the barrier for these types of users. These products also use far less data than other alternatives, which is a major benefit for users on very limited or extremely slow data connections.
The biggest difficulty with these products – at least the ones we’ve seen to date – is that there’s relatively little real artificial intelligence behind the scenes. Instead, there are basic text parsers that function an awful lot like a command-prompt interface. You might be able to ask an app “what can I spend this weekend,” but if you try a slight variation—say “how much cash do I have for Memorial Day,” you’ll probably get an error. In many cases, the user is back to memorizing commands, 1980s Microsoft DOS style. Doing so may be second nature to us geeks, but it’s a pretty ineffective way of interacting with most users.
A brilliant UI begs you to play with it, to try that weird looking button over there and see what happens; you can learn a lot by just exploring. But with text, you don’t even know what to try. You can flail around for a bit like in the old text-based adventure games, but that gets frustrating quickly. You can always ask the software to list the commands, but that feels like quite a step back, doesn’t it? None of this is fun or interesting, so most users won’t do it. If you have advanced features buried behind esoteric commands, it’s unlikely anyone will find them.
A chatbot-style interface works fantastically well when the tasks to be performed are very simple or—in the case of Digit for example—known ahead of time, and the messaging is mostly incoming. It doesn’t scale to advanced functionality very well, though.
On the bright side, this slimmed-down approach has allowed for a lot of interesting products to hit the market quickly. Client-side development is challenging, and building an elegant graphical user interface is a pretty substantial resource commitment. The bots are enabling companies to try out concepts much more rapidly and gauge consumer interest early. That’s great for innovation, but it also makes it harder to separate out the truly useful from the mostly trivial—both for consumers and investors.
The key question is whether we’ll see a return to more experience-rich products or whether AI will truly catch up and render speech and text parsing sufficient for complex tools. While still early, we are starting to see companies like Apple, Microsoft, and Amazon open up APIs into their voice-recognition products. If that continues, developers might be able to make use of the massive compute power behind these products to help navigate the vagaries of natural language expression. On the other hand, products like Google’s Instant Apps are allowing feature-rich experiences in an incredibly slimmed-down package, combining an elegant user experience with straightforward development and a thin client. My bet is on the latter winning this race, but if you disagree and think chatbots are taking over the world, let us know!