Conversational AI: How Messaging is Scaling Across the Entire Customer Journey

Alec Lazarescu
Bots + AI
Published in
5 min readApr 30, 2019
Crowd Panorama April 15th, 2019 Bots & AI NYC Event
Join thousands of other forward thinking leaders learning at Bots & AI events together at https://www.meetup.com/Bots-and-Artificial-Intelligence/

Our full house kicked off with Ryan Maynard, CEO of ChatDynamo (www.chatdynamo.com) — a digital marketing agency specializing in Facebook Ads. Ryan was working at Kik when he fell in love with the medium of messaging and hasn’t looked back since. He is an avid fan of basketball, sneakers, and chatbots (not necessarily in that order) and opened with a story about Google’s ad on TNT during the basketball to promote the Nike Adapt BB shoe.

The call to action was to ask your Google assistant to “talk to Nike”. 10 years ago the call to action would have been to “go to the Nike.com website” and 5 years ago to “download the Nike app”.

Ryan Maynard

Ryan walked the audience through 4 examples of the messaging medium disrupting advertising. Octane AI is disrupting traditional re-targeting methods (cookie-based ad pursuit of consumers that have abandoned shopping carts) with sending messages to the user via Facebook Messenger. Shopify merchants saw a 7–10% lift in sales using this approach. Messages sent via Facebook Messenger enjoy very high open rates (75%+) as compared to their traditional email counterparts.

Facebook Messenger’s “Click to Messenger” ads eliminate the need for re-targeting.

Ryan‘s top 5 practices for “Click to Messenger” ads:

1. Questions are the best way to initiate a conversation — make sure to ask a question in the ad and not leave users wondering how to

2. Always use welcome messages and avoid “Get Started” pages

3. Use Facebook Messenger ad units instead of Facebook Timeline ones… it eliminates the extra click for users and ad placement within the inbox yield in higher engagement rates — chatbot acquisition costs using Instagram are significantly higher

4. Refer clearly to “messenger” in the ad copy to reduce costs and reduce drop-offs. Using emojis can be attention-grabbing as well.

5. Facebook lookalike targeting (using Facebook’s algorithms to target ads to similar demographics of your existing audience) shows promising results

Next Ryan spoke about Beam (by Massively) which allows brands to enable chatbot conversations right within an ad banner on a website with yields nearly 2–10x higher engagement with ads.

Finally, Ryan introduced the audience to Cashbot.ai which is the first ad network for messaging based on the Chatfuel platform. Cashbot.ai makes it very easy to integrate ads into a bot without the time consuming process of finding and striking a deal with ad buyers. This is especially useful for bot makers who want to build fun bots without having an eCommerce or other direct monetization focus and still make money.

Shane Mac

The second speaker for the night was Shane Mac, founder of Assist and the new Chief Automation Officer at Conversocial (which acquired Assist this year). Shane authored a Medium post about Random Access Navigation (read here: https://medium.com/assist/theres-a-dozen-ways-to-order-a-coffee-why-do-dumb-bots-only-allow-one-27230542636d) that explains the importance of navigating customer intent effectively to create frictionless experiences. Part of that experience is considering and creating effective copy prompting for what you don’t know and need to ask the user to enter not just what you know.

The difference between the internet of the past and today is the consumer is in control. They can change their mind and type something else vs a web or app limiting the navigation path.

Shane showcased a few client case studies that he has been involved with through his work at Assist.

At www.1800flowers.com, Shane and his team created a live demo for Facebook’s F8 2016 conference, where for the first time a user was able to buy flowers using Facebook messenger. The messaging interface is personal, something most people use to interact with their friends. The future of e-commerce is that messaging will become the core channel.

Great brands of the future need to act like great friends

The next case study that Shane talked about was Sephora. Shane and team created an experience for Sephora using AR within messaging. Sephora users were able to buy items they tried, book an appointment for a makeover or talk to an agent without ever leaving the messaging interface. Sephora blended seamless integration with automation successfully which resulted in a 12% increase in engagement with the Sephora app. At Sephora they now have a focused conversation team — most of them evolved from roles in Mobile Product to Messaging.

Shane’s team also reduced the number of steps in the booking experience for Hyatt from 10–12 to 4–5 using messaging. This case study really drives home that messaging works really well with simple experiences to reduce friction and save time.

Shane ended with the following takeaways from his work at Assist:

1. Integrations are the killer app: payments, CRM and business APIs, data/AI offerings, and analytics greatly increase the user value of your bot

2. Errors are the intelligence — bot-makers should pay close attention to the these tremendous opportunities to improve the customer experience by capturing queries understood but not supported to offer advice on checking out that area in a retail store or an alternative that is available and for truly not understood offering a next step escalation to a human.

3. Using techniques like Random Access Navigation, focus on the users’ intent and allow them to freely navigate between intents instead of following scripted trees.

Shane’s mic drop quote of the day:

If everything is going to be commoditized by Amazon, the only thing you have as a brand is your relationship with your customers.

We loved hosting this dynamic event and hope you join us next month for more.

Join thousands of other forward thinking leaders learning at Bots & AI events together at https://www.meetup.com/Bots-and-Artificial-Intelligence/

Huge thanks to Rucha Gokhale for her work on this summary and Jason Valdina on photography! Want to join our team? E-mail humans@botsandai.com

--

--