Botness Montreal 2017 Debriefing
Botness first international event! It was really great to spend some quality times with the leading community in the bot/messaging/AI industry mostly from Montreal but also from Toronto and the US.
Big Companies Like Twitter, Slack or Microsoft were represented as many as various level Startups such as TroopsHQ, Snaptravel or Poncho The Weather Cat.
Broid attended the event and I’ll try to summarize for you the various talks & panels. But before that, let’s take the opportunity to thank all the organizers and especially Andy Mauro from Automat.ai. And last but not least, thanks to GSoft for hosting Botness, your offices are a proper wonder!
Introduction by Jean-Francois Gagné, CEO of Element.ai
If you haven’t heard of Element.ai, well they are probably the biggest AI solutions provider out there, turning the world’s most important Artificial Intelligence research into transformative business applications. Oh and they raised a whopping series A round of US$ 102 million back in June 2017.
So when talking about AI, Jean-François Gagné, CEO of Element.ai, is probably one of the most knowledgeable person on the subject. Some of his talking points:
- AI innovation having been focused on consumers and not really on traditional enterprises but now AI is eating up software
- Conversational UI is a great way to send feedback to AI systems
- UI is friction, the best system is a system that does not have a UI
But the part that I especially liked was when he talked about the fear/anxiety some people experience towards Artificial Intelligence (not aiming directly at Elon Musk). Jean-François basically said:
“The lack of transparency in AI is currently the biggest mistrust factor” — Jean-Francois Gagné
Ethics, accountability, transparency and privacy are probably the next big challenge. We’ll definitely try to tackle that in future blog posts.
Messaging Platforms Roundtable
This was a great panel, gathering four of the biggest messaging platforms out there:
- Ian Cairns — Group Product Manager at @ Twitter
- Sergio Silva — Director of Partner Success at @ Kik Interactive
- Bill Bliss — Partner Platform Architect @ Microsoft
- [Remote] Amir Shevat — Director of Developer Relations — @ Slack
- [Moderator] Mike Gozzo — Co-Founder & CTO @ Smooch.io
The discussions revolved obviously around messaging and the new Conversational UI paradigm. Here are my highlight selection from this panel:
- “Chatbot Backlash”: conversational UI is not the best UX for everything. If using a conversation (with or without chatbots) think if you are actually solving the problem in a better way that you would with a website or an app.
- “Initially the web sucked” declared Amir Shevat, meaning that the Chatbot hype may be down due to a lack in quality but we will definitely get there.
- Integrating ads into bots will certainly kill the industry as it would be like someone pushing you products in a one-on-one conversation #lame.
- A question from the audience: “What can you do make each platform work together because it’s really time consuming to integrate each API.” And here is the answer :)
Twitter Conversational Features for Bots Overview
Given Twitter unique social product, the real challenge with bots and Direct Messages is about bringing users back and forth between public and private conversations. Ian Cairns provided some answers to that challenge with a presentation of Twitter API’s upcoming features but also some insights about Twitter usage:
- 82% of all Twitter interactions are with brands
- The case study presented showed that users spent in average 5 minutes talking to the brand bot in Direct Messages
“ChatBots are not the Why they’re the How”— Ian Cairns
Driving Engagement in Bots
Moderated by Greg Leuch (Head of Product and R&D @ Poncho The Weather Cat), this panel brought some insights from Amanda Robinson — (Lead CUI Designer @ Automat.ai), Étienne Mérineau (Co-Founder @ Heydaychat) and Kevin Gauthier (Co-Founder and Design Lead @ Smooch.io).
The main outputs I gathered:
- Retention or engagement rates? There is no universal way to measure healthy bot engagement. You need to take the time to think about what problem you are solving, hence what success means for your bot you built.
- Don’t overlook quality-based metrics.
- What makes you a great conversationalist is actually being a great listener: Web truth: asking questions increases dropoff | Bot truth: users love to answer questions about themselves. If you ask interesting & relevant questions you’ll get your engagement and a ton of useful data!
Enterprise Customers Panel
Some interesting testimonies from bot users in the workplace. Frances Wilk from Breather spoke about their experience with Growbot which was a “human game changer” with 80% of employees actually using the bot!
Yellow Pages built “Flipper” an internal chat app to handle recurring daily tasks.
“All enterprise softwares will ultimately funnel into messaging” — Tom Hadfield (Co-Founder & CEO @ Message.io)
Raising a Series-A as a “Bot Startup”
This panel moderated by JS Cournoyer (Co-Founder of Element.ai and Partner @ Real Ventures) gathered 4 startups — not including Element.ai — that already raised several millions in seed and/or series-A.
During the conversation, everyone seemed to agree that the “bot hype” was quite over (for now) and now you can’t really raise “easily” seed money for a bot startup unless you have a team of successful serial entrepreneurs or that provide a real unfair advantage.
“The question is not if conversational interface will take over it’s when” — Tom Hadfield
On day 2, David Nault from iNovia Capital took a moment to speak about teens in distress and the fact that Messaging and Bots can definitely provide some help regarding that sad and overlooked issue.
If you wan to help, he is currently building a task force, just contact him with your offer.
Consumer Customer Panel
We heard the testimonies from both BNC and Loreal Canada with Martin Aubut. For the latter, the opportunity of using Conversational Marketing presented itself when they opened Messenger on Loreal Facebook Page. They received hundreds of messages per day! The dialog with their customers was open, they just had to seize the opportunity!
“We need to create value to fuel demand and mature bot usage, we’re not there yet”
How SnapTravel is building the leading eCommerce Bot in Travel
That was a great presentation from Henry Shi (Co-Founder & CTO @ SnapTravel).
Just like so many “bot” startups in 2016, SnapTravel actually started as a “human behind the wheel” agent. But users loved the experience so much that they could scale with an actual bot.
Advice freebies:
- Leverage the right medium: Chatbot, Human Agent and Webviews at the right time.
- “Test, test, test” Snaptravel is running 20 experiment at any moments.
Automatic Evaluation of Chatbots
I’ll be brutally honest: I can’t give any takeout from that presentation by Joelle Pineau from McGill University. First, because I didn’t get 100% of it and secondly because that was rather very (very) technical. But it was pretty interesting to learn about the Reasoning & Learning lab, what they were working on and how they tackle AI chalenges.
Chatbots and conversational UX, an investor perspective
Jason Brenier from Georgian Partners did a beautiful work of explaining the Principles of Conversational Business gathered into four main segments:
- Purpose
- Data
- Design
- Integrity
“A Chatbot is a Brand ambassador, so you have to make sure it represents your company well”
Here you go! If you made it here, thanks for reading and if you did not, well then I guess it doesn’t really matter.
All in all it was a great event with awesome industry people. Let me try to conclude with my own analysis of these two days of talks.
Even if it’s currently going downhill, the chatbot hype showed us that the Conversational UI is the future of eCommerce/Marketing/Customer Support but it has yet to mature. That will only happen when value will be provided to the users. How can we do that? You don’t absolutely need a fancy AIish chatbot — Users are as keen as talking to a bot as to a human. But it all starts with a conversation right? So Funnel your various messaging channels into a single place, use human to answer those messages, use enriched data to improve response workflows and then you’ll be able to automatize some tasks.
Bots are basically like humans, they need to grow, learn, fail and succeed. You wouldn’t dress a 6 year-old child like a maestro for its first performance, right? That’s the same with your bot: users will expect your bot to be a perfect professional but he won’t, he will be a baby making its first step into the world.
But don’t worry, today’s baby is tomorrow’s adult.
Cyril “Pepito”
You shouldn’t spend time integrating messaging APIs. You should focus on your core product. With Broid API, get instant access to all the major messaging channels as well as our own Web Messenger powered by the best conversational features available.
Broid is also open-source based, check out our GitHub Repositories.
Built with ♥︎ in Montréal, Canada
