A More Forgivable Bot?

Vesna Jocic
Jul 10, 2017 · 3 min read

Earlier this year, the number of bots on Facebook Messenger surpassed 100,000. If you haven’t yet had the chance (or inclination) to engage with one of these “intelligent” bots — on Messenger, Kik, Slack, WeChat, or any number of other platforms — chances are you’ve heard complaints from people who have.

Within two days of Facebook launching bots on Messenger, Gizmodo had declared them “frustrating and useless”. While the functionality of some bots has improved since then, for the most part, the experience of chatting with a bot still leaves most users less than satisfied. In fact, Messenger veep David Marcus insisted recently: “We never called them chatbots. We called them bots.” Apparently, “People took it too literally in the first three months that the future is going to be conversational.” Sorry we were so literal Dave!

In a lot of cases, the frustration people feel while interacting with a bot stems from the fact that it’s designed to converse with us as if it’s a living, breathing human being, without really having the intelligence to do so in
a way that’s convincing and enjoyable. And even as a bot’s ability to mimic human conversational patterns improves, our recognition of the ways in which it still falls short grows. As Kaveh Waddell, Staff Writer at the Atlantic argues, “these bots have entered their own version of the uncanny valley.
As interacting with them approaches the experience of talking with another human, their robot-ness becomes accentuated.”

What if bots weren’t trying to be human?

I was recently tasked with creating a landing page to promote an internal corporate chatbot. The problem: like most other chatbots, it couldn’t actually do very much. It could help users find someone’s email address or book a meeting room—in theory, though I never got it to work — but that was about all it could do. How was I going to convince users to adopt a tool that wasn’t very helpful (yet)? After mulling over the problem for a couple of days, it occurred to me: If you can’t be helpful, be cute.

In this case, “be cute” meant “be a dog”. The chatbot was presented to users as a puppy who was trying to be helpful but only knew a few tricks. When the dog wasn’t able to understand a user’s request, it offered to send a funny dog gif instead. It’s far from a perfect solution, but freeing the chatbot of the expectations that we place on humanlike entities seemed to make users more forgiving of the bots obvious limitations.

Perhaps one of the most well known (and liked) chatbots with an animal avatar is Poncho, the weather forecasting cat. Poncho is designed to “send fun, personalized weather forecasts every morning.” But that’s not what makes Poncho so popular. If you ask Poncho to tell you a joke, he will.
It might not be an especially funny joke, but he obliges. And if you laugh, Poncho laughs too.

“Instead of simply rephrasing questions into statements and statements into questions, Poncho responds to input and shifts topics in meaningful and interesting ways,” writes Senior Editor at Poncho, Ashley D’Arcy. “Poncho works because when you go off topic, Poncho goes there with you.” If you make an inappropriate comment, for example, Poncho will ask you to apologize. If you get flirty, he’ll ask you out on a date. It’s fun, but I can’t help but wonder whether people would find a human version of Poncho
as engaging. Anyone in the mood for some A/B testing?

Vesna Jocic

Written by

Designer | Developer | Artist | Night Owl | Tennis Addict

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade