What Chatbot Design Taught Me About Empathy
And how it can help you design for what matters most
Ever since I began my journey as a UX Designer, I had always designed for digital experiences.
This past week, I learned a lot about designing for conversation.
I came to realize that in our increasingly digital world, we are constantly seeing shifts and improvements in how we experience human-to-human communication. However, something we all take for granted is the art of designing for human-to-AI communication.
If you think about it, most of the technology you use depends on AI to provide you with the in-context information you need. The same is true with chatbots.
Chatbots are everywhere. Since the term “chatterbot” was coined in 1994 by Michael Mauldin, chatbots have become an important part of our daily lives.
Thanks to chat-based AI, we’ve been able to schedule tasks, set automated responses to customer questions, ask the status of the time/weather, and so much more. We’ve practically automated the majority of tasks that were once written on paper.
Besides Paying a High Salary, It Demands Unique Skills
What if I told you that the average bot developer makes $106,000 a year? And that the industry is expected to be worth 1.25 billion by 2025?
I bet you’re interested now!
(Don’t feel bad. That would catch anyone’s attention.)
This is because despite the high pay, you definitely must have a genuine interest in this field. Knowing the ins and outs of the human context — how a person would approach a particular situation — is essential to designing a chatbot from a human-centered point of view.
When took to developing my first fully-functional e-commerce chatbot, I had to consider where the interaction could potentially happen. Since the client had a Facebook business page that was fairly active and growing, I decided that I would start there.
For this particular project, I chose ChatFuel, a free chatbot development tool designed for Facebook Messenger. Despite having a bit of a learning curve, the software becomes easier to use with time and effort. With enough practice, even the most inexperienced user can craft an enthusiastic chatbot with just a little imagination.
But knowing the software itself isn’t enough — even the most technically literate creatives are reduced to nothing without abiding by the laws of human-centered design processes. Here are 3 lessons I learned from these moments of trial and error.
Lesson 1: To Design For The Customer, Observe The Customer
Observing the customer meant viewing recent fashion industry articles, and even reading books on sales language. My client’s particular industry was fashion, and this required that I approach the chatbot design process in a way that answered three important questions:
- What products or services would the average online fashion customer want?
- What problems could shoppers have with the client’s business that this chatbot could address?
- What moments of delight can I include to bring life to the conversation experience?
All of these questions provided a strong foundation upon which I was able to construct a fully functional messenger designed to solve problems for potential customers. It was during this phase that I first discovered that customers shopping online for fashion products, especially the fashion products my client Meraki Clothier provided, would likely not have a familiar grasp of what Meraki Clothier’s brand unique from others.
Luckily, my client and I had just come up with a new idea: a content creator platform designed to help promote the work of emerging digital artists on the online store’s apparel and social media.
As I explored the potential opportunities, I especially wanted to consider how this platform could add value to the overall user experience.
Keep in mind that I had not begun building the chatbot yet, for only after considering these value propositions could I begin design work in Chatfuel.
Lesson 2: Conversations Should End When The Customer Wants Them To
Chatbots are often associated the annoying, intrusive predecessors. For example, you had Clippy, which was an AI tool often remembered for how it would open each time you launched a Microsoft program on your computer.
Clippy was designed to help users find solutions to issues based on their specific actions, which sounds like great technology. However, this AI tool failed as a key feature because it violated many social norm. And as time went on, people noticed it more.
Every designer likes to think that if their solution solves a problem of the present, it’s practically foolproof for times ahead.
When designing Meraki Bot from a human-centered perspective, I considered how this AI could give users the opportunity to immediately end the conversation if they wanted. I wanted to make sure each element of the conversation felt seamless and personal.
To do this, I thought in reverse and considered the moments in which a potential customer could either 1) Re-start the conversation cycle, or 2) End the cycle altogether.
For example, when designing the initial conversation, I added the option “I’m fine” at the bottom so the user can immediately end the conversation or request chat with a human.
Considering these factors greatly improved the quality how I applied research and design principles. It wouldn’t be long before I began thinking of new ways that this chatbot could serve customers and provide unique moments of delight to their everyday shopping experience.
Lesson 3: Allow Space For Human-To-Human Interaction When It’s Needed
Every designer likes to think that if their solution solves a problem of the present, it’s practically foolproof for times ahead.
However, times change — and as a result, user behavior does, too.
Factoring this in, I considered the moments where a user may want to immediately get in touch with a live person, and request customer service that requires multiple steps to handle. Designing for these moments was tricky at first, because to build a strong conversation, I had to understand how the chatbot could advise the user to solve the issue themselves if possible.
I thought of the scenarios where human assistance would be best suited to handle customer requests. After all, if the chatbot found that it could not solve the customer’s problem, then it should direct them to a human who can.
This approach helped me think beyond the idea of the chatbot being a tool — it helped me better understand why the chatbot was one of many parts of a fully integrated whole. This is why Meraki Bot was built on the philosophy that automation was needed to make the customers’ shopping experience easier and more robust.
What You Can Take From This
Designing from a different perspective than what you are accustomed to everyday is a great way to build appreciation for the many fields in design that are out there.
Best part is, you don’t have to be an expert to start!
When building this chatbot for Facebook messenger, I didn’t go into it expecting anything grand to come out of it. I simply took to the act of building a product that a business needed based on the unique and changing values of its customers.
So you continue to move forward in your design career, think of the ways you can learn to design for new and non-mainstream user experiences. It’s worth a try, and you’ll have learned something in the end.
