Thanks to Women Who Code, I got to attend this year’s #SuperBot17, the premier conference for AI, Chatbots, and Virtual Assistants at Automattic, organized by friends at dashbot. I arrived just in time for the Engagement & User Experience session mid-morning, and while I’d missed the keynote altogether due to some delays in getting there (hello, Bay Area traffic), I learned so much throughout the day and met some interesting folks whom, like me, love to geek out on emerging technology. I had written my first medium article earlier in the week about my own experience in building a chatbot for non-profit use case, so I was pretty excited to learn from the “super bot” experts.
Here’s a breakdown of each session and key insights I gained from the conference:
Training bots to have higher EQ (emotional quotient) was the common thread across all three presentations from Poncho, Pullstring, and Assist. Success of bots is measured by engagement and user retention, and the best way to do so is to give users a reason to stay once they engage. In contrast to emails, chatbots currently boast a 97% open rate, according to Arte Merritt at Dashbot, which alone gives more reasons for organizations to take a plunge in the chatbot space. As far as improving user experience, Stephane Nguyen from Assist says, it’s the subtleties that count:
Copy is important, but we are driven by friction and reducing that friction. — Stephane Nguyen, Assist.io
As an aside, one pet peeve of mine as a designer has been that the creative capital always seems to be undervalued. In UX, the best design is always one that is invisible, and here it’s no exception.
The panelists who work with household names like The Today Show, AOL, Inc., CNN, MTV, Kia motors, and Starbucks, to name a few compared notes and shared their insights from helping big brands reach new customers through chatbots. It was interesting to hear about their successes (Reply.ai partnered with Kia motors to reach out to 8,000+ new buyers with their chatbot) and struggles (no one wants to dedicate resources for bots yet so you just have to make it on the side until they show interest).
bots should have a clear purpose and character but they are not yet at the maturity level to disguise themselves as humans. — Omar Pera, Reply.ai
When asked about the future of bots, one panelist predicted:
Voice is the new interaction. Bots will be the new browser. — Drew Lesicko, AOL
It’s a pretty sweeping statement, but it was enough to make people look up from their smartphones. Ok, so how do we monetize that? Omar Pera from Reply.ai asked the audience to proceed with caution:
I’m worried that advertisements will bring down the bot experience. Think about the users’ perspectives first. — Omar Pera, Reply.ai
I asked: “Can you tell us about designing bots with ethics? Everyone here seems so focused on monetizing bots, but how do you make sure you build good bots as part of your brand strategy?”, and here are the responses I got:
We want to set the right level of expectations, because the last thing we want to do is to anger our users. We’re starting to think about our users first. — Andrew Pinzler, TODAY show
And here’s my favorite response:
Today, I strongly suggest to not mimic a human. Make your bots with character, personalize it, and put a lot of content into it, but try to set your objective first. Think about what this bot is going to do, implicitly or explicitly, but say that it’s a bot. Just imagine the worst case scenario when your user finds out they were talking to a bot rather than a human. You have lost their trust forever. — Omar Pera, Reply.ai
What sets bots apart from web/mobile is that conversational UI is richer.
Dennis Yang from dashbot explained that mining data with chatbots using proprietary analytics platform is near impossible, yet he found that users share a lot more data with chatbots than through traditional platforms. If you want to read more about their fun-alytics, Arte Merritt has just published this article via Medium on “Common (and uncommon) messages sent to chatbots.
Incidentally, Amazon Show was launched on the same day as the #SuperBot17 conference. The thing looks like it was 3D printed right out of the Jetsons and it’s a matter of time before someone starts selling hand-knitted screen covers for its early adopters on Etsy. The smart home voice assistance market had ballooned in the last two years, with Amazon Echo taking the lead with yesterday’s Fortune report confirming it has 70% of the U.S. market. Vera Tzoneva from Google Home, Amazon’s closest competitor, said this:
20% of mobile searches are now via voice.
Since, of course, Google still dominates as the nation’s preferred search engine. Companies like Amazon, IBM and Google are investing heavily into Voice AI development environment, because in order to stand out in the market, quality matters:
Everything goes back to the quality of your voice product. Discovery, engagement, and retention depend on quality.
The panelists agreed that it’s unlikely that voice will take over as the singular medium, but rather, it will supplement existing services and other emerging forms of AI assistance technology.
Best Prototyping Tools: Protobot, BotFrame, Walkie, BotSoceity
Best Code Frameworks: Twilio, Slapp, Botkit, BotFramework, SuperScript
We also had a demo from Ben Brown at Botkit, which boasts an average 1 download per minute for its open-source toolkit he created.
It’s not surprising that VCs are betting on new technologies and looking for ways to monetize it, but the consensus seems to be that they are hedging their bets by looking at multiple bot platforms, whether it’s voice, hardware, or chat.
Lastly, I thought this quote from Dennis at dashbot was pretty memorable.
Conversation without listening is just talking.