Chatbot Design – Not all chatbots are created equal

There are now thousands of chatbots out there. Major consumer brands like Louis Vuitton, Domino’s,  Air France, and L’Oréal have started investing in conversational marketing. But for all the buzz surrounding chatbots, not all of them are created equally. Some deliver brilliant experiences while others are just not up to scratch.


Chatbot design - Brands investing in Conversational Marketing


Chatbot design - Beauty Brands investing in Conversational Marketing


An effective chatbot enhances the customer-brand relationship, while a poorly designed one can harm it.  So how do you tell a relationship building chatbot from a relationship killer?  My colleague Amanda Robinson recently tackled that subject as a featured guest on Jon Prial’s Impact podcast. Amanda is an expert in chatbot design. She’s been working on conversational technologies since the 1990s and worked for Fortune 500 companies before joining Automat where she works with our customers on conversational marketing strategy, interaction and design.

If you have time, the entire podcast is well worth a listen.  But if you’re pressed, here are the top takeaways from Amanda’s brilliant interview.



Messaging First – The key to Chatbot Design

Chatbots are interactive and a great design experience takes full advantage of that.  Yet some chatbots are still narrowly conceived with email-like, one-off interactions in mind.  Think walls of email-styled text that burden users with too much content or clunky questionnaires that pressure customers to narrow themselves into a category. Another classic example is using a promotional deal to lure consumers to a product or brand.  You may make a sale, but when you “hook” customers with a discount, you’re setting them up to only expect deals from that channel, and that’s a missed opportunity.

Chatbots can deliver so much more.  The key is to put messaging at the forefront of design decisions. At Automat, we call this “messaging first” and it opens up all sorts of opportunities by creating truly great experiences.

Take text for example. A “messaging first” designed chatbot takes full advantage of messaging interactivity by speaking with consumers in very short bursts, as opposed to hitting them with email-styled walls of text. This creates a back-and-forth exchange that is similar to how users may talk to their friends and family. The result: consumers get an experience that is lightweight, organic, and enjoyable while brands get to know their users, and then build on that knowledge to maximize customer lifetime value.


The Beauty of the Soft Sell

It’s really easy to spot a poorly designed chatbot.  People don’t chat with it. There are many reasons why this happens, but the biggest reason in our experience is that the chatbot is asking users to reveal too much up front without explaining how they’ll benefit from sharing that information. In other words, a chatbot with a hard sell is an invitation for customers to disengage.

For example, if a chatbot asks something too quickly, such as the user’s age, without establishing why it’s asking such a personal question, it may give the user pause. They might even wonder if that data is being harvested simply to sell on to a third party. That’s not to say that asking lots of questions is necessarily negative.  But when the conversation feels forced because a chatbot lacks empathy, it can backfire.  


Messaging: Where Chatbots come into their own

Web chat, SMS and email have their virtues, but messaging is where chatbots really come into their own. Unlike SMS, messaging can help brands create a fully interactive experience with images, videos, sound, carousels and clickable buttons. With web chat, conversations can disappear as soon as a user closes their browser. Messaging, on the other hand, establishes a persistent thread with rich analytics that deepens a customer-brand relationship over time. And unlike email, which organizes conversations by topics and threads, messaging categorizes conversations by people, which further enhances the deep relationship.

Of course, messaging platforms can differ wildly, and brands need to research which ones are right for them. But once they identify the best platforms for achieving their business goals, brands can design chatbots to build multi-turn, deep interactions with customers via messaging. As the data flows back, brands can see what’s performing well and what needs to be improved, and then adjust their chatbot design accordingly.


Building a Chatbot that’s right for your Company

A great chatbot design starts with brands identifying what they hope to achieve with it. For instance, some brands may want to migrate an audience from email to a messaging platform in order to offer them something unique. Other brands may want to attract new users to the experience. Whatever the goals may be, once they’re established, companies can frame three core metrics to assess whether the chatbot is helping them achieve those aims: usage & engagement, conversions, and satisfaction.

Usage and engagement metrics measure things like total conversations, total messages, the average duration of a conversation, or the average number of turns per conversation. Conversion metrics help measure business goals, like the number of rich profiles gathered. Satisfaction metrics go even deeper and measure whether the users’ goals are being met.  Think thumbs up or thumbs down. Or comments. And don’t forget that engagement itself is a proxy for how satisfied users are with a chatbot experience.


Ready for more?

Those are my top takeaways from this podcast, but if you want to go deeper, I highly recommend you listen to Amanda’s complete interview.   And remember, with Automat, it’s easier than ever to get started on your conversational marketing journey. So let’s chat!