A Quick Look Back at the 2017 Chatbot Conference

  • BenJamin Morin
  • September 25, 2017

There are a few quick observations to be made about the 2017 Chatbot Conference. First, San Francisco in September was beautiful. Second, the Chatbot community are a passionate group.

The one-day conference was chock full of nuggets shared from those paving the way. There were presentations from Slack, panels of success stories and marketing ideas.  People talking about the future, the possibilities and some of the amazing technological breakthroughs that have made the current state of Chatbots possible.


Here are a few takeaways from the conference that I felt need to be shared:

Do one thing and do it really well.

This has been the driving force behind many mobile applications over the past few years.  I recall opening my American Airlines Mobile app and seeing a game of Sudoku inside.  I assume that this was something they thought users might do while waiting at an airport.  That still makes little to no sense.  I have over 300 apps on my iPhone.  If I want to play Sudoku, I will go to the gorgeously designed Sudoku app I already have installed rather than the mediocre game inside the airline’s app.  This digression is really to point out that your chatbot should not be trying to do more than you intended it to do.

Users might misunderstand what your chatbot can and can’t do.  By focusing on one thing, you keep the conversation focused on the topic at hand.  From a creator perspective, it also keeps your team focused on providing the best experience without distraction.  Think of a chatbot as a conversation.  You wouldn’t stop talking to a customer about your products and services mid conversation to discuss pop culture trivia, so don’t build a pop culture trivia conversation inside your chatbot.


Bring your chatbot to as many platforms as makes sense but don’t just copy the experience.  Depending on the type of business and target demographic, it might make sense to have an application on text, voice and Facebook messenger for example.  The key thing to remember is that the user experience is different in each case.  Where the text and voice might use the same dialog, sometimes inserting emojis into a text improves the experience whereas trying to express the emoji in a voice experience is awkward and downright strange.  Take it even further, Facebook has a much richer UI experience that, if treated exactly as the SMS experience, may see a drastic loss of users.

Consider your audience

Along the same lines as decentralizing, you should know your audience.  If your primary audience is Men ages 40 and over, you may not use as many emojis as you would if your audience is under 30.  You also have to consider regions.  Certain dialects of the English language use certain key words.  That’s why, if you are working in a specific region, ensure your bot is familiar with the language.  If you are larger, make sure the dialect you work with is generic enough to work across all of your regions.  Don’t even get me started on language localization.

Everyone should have a chatbot and they will all be ugly

Let’s face it.  Right now, most of the bots out there are not very pretty.  They are in fact messy and need lots of attention.  This is why I like to think of this as the infancy of the chatbot era.  In order to have that fully realized bot you want, you have to start now.  It’s going to take time and nurturing, but like a child, given the right stimulus, it will become something you can be proud of.  Getting in now means going through the same growing pains that everyone else is experiencing.  In five years, while you are finally starting your first bot, everyone else will have been through it and you may have missed the boat.  Consider this your dose of peer pressure.  You are welcome.

Don’t expect it to change your world overnight

I would ask many of the presenters the same question after their talks.  How do you measure success?  I heard a variety of answers but nobody had “the answer”.  This is where it gets tricky.  If you start a chatbot today, don’t expect all of your traffic to go through it tomorrow.  I think my favorite answer to this came during an afternoon panel.  I wish I could remember who said it, to give them the credit they are due.  Their response was monitoring “containment rate”.  This is the rate at which bot conversations must convert to live person conversations.  So rather than focus on the effectiveness of having a bot vs not having a bot, they are measuring the quality of the conversations.  For now, that is a great perspective and open minded enough to drive a business to success over time.

There are so many other great topics that were discussed.  From the changing world of UX, further discussions on multi-modal trusted AI personalities (think chat, voice & avatar videos) to Bots rights (bots are people too).

Final thoughts

On top of the takeaways from the conference, there were two other great pieces that came from it.  The “Game of Bots” was a bot written especially for the event.  It gamified the conference, brought focus and additional information to each panel and created a fun way to involve the audience.  I also can’t complain as I won the game and came home with an Amazon Echo Show.

The final take away, and something I recommend everyone do, was the great networking opportunity.  The chance to meet and engage with others building bots to discuss challenges and plans allows a creator so much more of an advantage than those working in a vacuum.  While there are plenty of people working on bots, the community can always use more.  Competition breeds ingenuity.

Go.  Make your bots.  Make them even if they are stupid and ugly.  Raise them and mold them and teach them to do amazing things.  Just don’t let them take over the world.  I live there.