What we learned from Botness, a conference for chat bot ecosystem

Originally featured on CareerLark Blog. CareerLark is a Slack bot that helps you get micro-feedback from your coworkers. We’re hiring! Follow us on Twitter and LinkedIn and like us on Facebook to receive updates on new posts and developments.

The CareerLark team attended Botness, a conference organized by leading messaging platforms and bot developers for the purposes of “shaping the evolving ecosystem to support innovation and open collaboration.” During these two days, platform makers, tools developers, and bot builders got together for lightning talks, unconference discussions, and impromptu collaborations.

Much has already been written on the topic: for starters, check out Chris Messina’s piece on Conversational Commerce, Sarah Guo’s series onConversational Economy, and VentureBeat’s Botness coverage. Against this backdrop, here are some high-level themes that struck us.

1. Sentiment summarized: nervous excitement

Let’s unpack that, starting with the excitement.

One of the first things that we noticed was a palpable sense of excitement around the enormous potential of bots. Matt Marshall of VentureBeat listed the major three forces that are creating this buzz:

1. AI. Only now AI is on the verge of a breakthrough for creating useful, scalable products.
2. 2–3B people in chat apps. The market size of chat app users have grown exponentially in recent years to the point where chatbots can become a mass-market mainstream stay.
3. conversational UI. Chat UI is becoming the de facto interaction on mobile, simply because we spend increasing amounts of time in messaging apps. Chatbots are poised to execute tasks on behalf of users using this emerging paradigm.

Or as Chris Messina referenced it in his intro talk:

So why the nervousness?

AI is one of those fields that’s perennially been on the verge of a breakthrough. There were attendees that have been working on AI/NLP/ML for decades. There have been cautionary tales of early bot successes that have not translated into long-lasting success, the poster child beingSmarterChild. It was amazing and humbling to hear the story of the ups and downs of SmarterChild from the man and creator himself Robert Hoffer.

It is true that 2–3B people are using messaging apps today and as Chris Messina noted for the younger generation, conversational UI is the de facto way of experiencing technology. But this doesn’t mean that chatbots using conversational UI is the best way to solve problems; in fact, a nice old-fashioned GUI might be the best solution as discussed by Dan Grover here.

There was incredible hope that this time, finally, we are riding the true wave…but with the nagging doubt of what if? What if technology can’t quite deliver on consumer promises? What if reality doesn’t quite meet the hype? What if chatbots went the way of QR codes vs. mobile apps? How do we manage consumer expectations such that they don’t abandon this as yet another fleeting novelty? These were some of the existential questions that were debated at Botness.

2. We have a thriving and growing ecosystem of bot makers

It became clear to us that we live in a world where an increasing number of startups are tackling infrastructural and tools components of building bots. Ben Brown summed up the philosophy of making products for bot makers:

Startups such as ManyChat threw up some impressive adoption numbers to the tune of hundreds of thousands of bots created and Mike shared his thoughts on distribution and use cases.

A morning session of framework providers made it clear that startups are taking different and creative approaches to solving some of the hairiest problems in bot creation. It was encouraging to see that even the least tech-savvy could now stitch together existing tools and create their own bots!

3. UX needs to improve across the board

Much has been written of disappointing initial bot experiences (like here). That sentiment was readily echoed by bot makers as we acknowledged that tech and design have a long way to go in helping us create a world-class “wow” user experience.

Here are our thoughts on why:

1. AI still needs work. There is starting to be an increasing number of tools out there, but no one has nailed the right combo of NLP + ML that bot makers can use as turnkey solutions. In fact, in an unconference session on AI, most of us bot makers raised our hands when asked the question of whether we were building our own solutions or using a service such as wit.ai or api.ai when converting chat inputs to actionable data. This simple question showed that despite the significant advances we’ve had so far, we still have a long way to go.
2. We don’t know how to best design for conversational UX. This is a whole new paradigm and that means design best practices are completely different than what we know. Botness did a great job of starting best practices sharing on this; for example, Omar from Sequelencouraged us to think about bots as personas. Ask yourselves: would you rather get news from AndersonCooperBot, or CNNBot?
3. Many bots are side projects or from hobbyists. One byproduct of making it easy to stitch together tools for bot making is that it is easy to produce a relatively low-quality bot. And while this is great for the ecosystem, we also need to invest in infrastructure to help bot companies build high-quality bots that deliver significant value to our users and that we can monetize.

4. We have a lot more to gain from collaborating vs. competing

One of the best parts of Botness was the spirit of collaboration, even between competitors. It was great to see that we all acknowledged that a rising tide lifts all boats, and given how early we are in developing this ecosystem it’s in our best interest to share and collaborate.

This manifested itself into Unconference sessions such as Startups Meet Platforms (twice!), where we discussed what platforms can do collectively to help the developers.

A number of action items came out of that meeting, and we look forward to seeing the progress in the coming months!

5. And finally…building a great bot is really REALLY freaking hard

Have we mentioned that building a world-class bot is incredibly difficult? :) One of the best parts of the conference was that we learned that we are not alone. We had a number of “Oh no way, you have that exact problem too?” moments, which made us realize that we should be more proactive in cultivating a community of learning together.

We have yet to answer the million-dollar question: what chatbot will be the first to get to 1 MM MAU? And how will it monetize (if at all :)? We are in early days of bot business models and discovery, and it’s exciting times to be a maker in the bot ecosystem.

Thank you Lili, Ben, Brady, Amir, Ivar, Chris, Mimi, and John for organizing this conference, we look forward to future gatherings to come!

Enjoyed the article? Click the ❤ below to recommend it to other interested readers!