The 3 Things Chatbot Developers Should Know

by Brandon Gredler and Chris Wooster

Chatbots are cropping up at big brands across many different industries from the media to hospitality. These digital applications help businesses serve customers by automatically responding to online requests, helping people book hotel rooms, or finding store locations. They’re called chatbots because they replace real people on the other end of a chat line. Typically, these digital avatars are set up to respond to specific questions with specific answers, but more are being created with artificial intelligence and machine learning software so they can adapt to unplanned conversations and respond in a more conversational tone.

For all the advancements being made in the category, there are still three foundational elements all developers should consider before creating their next chatbot.  

Have a clear purpose
It can be easy to succumb to momentum of bright shiny objects, but creating a chatbot just because everyone else has one is no good reason to do it. Ask yourself, why are you building this? What value will it bring? Understanding the answers to these questions will define the purpose. For example, do you want the chatbot to be purely educational or will it expedite a sales process? Or is its sole purpose to reduce call center overhead? By setting the goal of the entire project, you will have focus and results to measure.

Understand the platform
There are numerous publishing platforms on the market—whether it’s Slack, Facebook Messenger, SMS, or WhatsApp—so developers need to understand those platforms just as much as they need to understand what they’re trying to accomplish. The way a platform operates will become just as important to completed design as the conversational flow of the chatbot. Different mediums and interfaces have various mechanisms for conversing and completing tasks. Based on the overall purpose of your bot and the desired user set, different channels are going to be better for you to use than others. Once the platform is decided upon, the team can start layering in things like personality, an element that will also be formed by the unique features and functionalities of the selected platform.

Don’t set it and forget it: chatbots evolve
Before you launch a bot project, remember that these are not set it and forget it applications. They require monitoring and optimizing. By consistently listening to and observing the chatbot, you will be able to understand new use cases, various language used, new utterances, expanding customer expectations, and more. These are all crucial to know so that you can evolve the chatbot over time.

Designing for voice requires a new approach
In general, developers should treat chatbots like an entirely new discipline. You’re now monitoring for things like fall back rates and rates of confusion. Not only do teams and workflows shift, but we have to rethink KPIs. For example, it’s typically a good thing if a customer spends an extended amount of time on your site. But if visitors spend a long time conversing with your bot, it may be a sign that your conversational flow isn’t clear or concise enough.

Also, when designing and developing a chatbot avoid overly rigid frameworks and processes. This is a relatively new design expression that is rapidly evolving. With chatbots (and voice-based designs in general), a whole new set of design principles and practices need to be defined. Tried and true user journeys, UX patterns, from the web or mobile do not always apply. For example, with the shift of workflows, you’ll need to rethink teams. Your creative director is now your conversational designer. Your animator is now your linguist. And so on. Instead of designing for pixels, we’re designing for conversations.

Brandon Gredler
Chris Wooster

More From These Authors

Related Thinking

End of the Week


End of the Week

Brain Manipulation and Shopify Flips the Script

Trends on Tap: CES


Trends on Tap: CES

Brandon Gredler, Director of T3 Labs, shares some of T3's favorite tech from CES 2020.

Starbucks Sleighs with New Instagram AR Lens


Starbucks Sleighs with New Instagram AR Lens

After 22 years of releasing newly-designed holiday cups each year, the brand decided to add a new element in 2019 by creating an AR feature on Instagram Stories for the cups (think Snapchat filters). If a user has one of Starbucks’ holiday cups, the AR effect will detect the cup and create a neat animated effect around it unique to each design.