Wizeline UX Designer Florian Lissot and Chatbot Conversation Designer Diana Lee discuss how they work together to create a user-centric chatbot experience. Together, this unlikely duo solves specific pain points and delivers quantifiable goals.
In conversation with Florian Lissot and Diana Lee
For products like chatbots, copy and design must work together to create the best user experience. Working with a prototype that evolves to meet user needs means that it’s not a streamlined process. It’s never as simple as using a placeholder for copy, then handing off to engineering for implementation. The close relationship between conversation design, UX design, and engineering is essential and somewhat unique to building custom chatbots.
Q: What is the value of having two completely different types of design involved in the process?
The beauty of combining two people with different backgrounds and experiences is that they bring different perspectives to determine what the minimum viable product (MVP) for a client should be. The UX designer must consider the visual elements familiar to smartphone users, such as spacing and motion effects, while the conversation designer must focus on using language and tone that matches users’ style of speaking. Having these two perspectives results in less error and a more natural user experience.
F: Yes, we’ve reviewed our process and determined that discussing each other’s feedback and aligning earlier allows us to create a better product faster.
For example, we recently built a flow for users to schedule a meeting time using the bot. We agreed that we needed to ask the user for two pieces of information—time and duration. Based on this info, the bot suggests rooms available during that time slot.
When I wrote sample copy to request the time and duration in one question, Diana suggested we break it up into two separate questions so the user would only have to think about one thing at a time. We discussed back and forth how best to break up the questions. In the end, we followed the golden rule: users should be able to reach the end-goal in as few steps as possible.
Q: How does user testing inform your decisions?
D: Testing with the intended users informs us of tweaks, and sometimes entire redesigns, that need to be made for the first delivery. As a conversation designer, it tells me what text or visuals (images, GIFs, videos) need to be adjusted to make it crystal-clear what the user can accomplish with the chatbot. The goal is to have as few fallback messages (a message with generic menu options) as possible.
This requires me to adjust the wording or change the order of messages so the timing and cadence feel natural and helps users achieve the task quickly.
Fatigue is also common for users. My job is to ensure that text fatigue doesn’t happen with our chatbots. As UX Designer Eunji Seo says, “Don’t make users go TL;DR.” Generally, anything more than three lines of text is too long.
Q: Florian, which UX principles apply to chatbots and which ones do not?
F: There are many UX principles that apply to chatbots, but the truth is we can’t rely on these alone for customized bots. Each bot has different user personas and pain points to solve. It is crucial that we test the bot with the people it is designed for.=
However, there are some generally applicable rules.
If there are 3 call-to-actions in a card element, the user’s eyes will be drawn to the last button. We also recognize that users must be able to exit and enter flows easily by clicking on another button. We try to use actionable copy in each button.
One thumb, one eyeball
This is a concept coined by Google’s Product Director, Luke Wroblewski. Typically, mobile apps are a “one thumb, one eyeball” experience. In a distracted environment, mobile users use one hand to perform tasks within a short attention span. We must design visuals and text to accommodate this phenomenon.
A good litmus test for this is “Can users perform a certain number of tasks with just one hand in under 60 seconds?” This, of course, is not a hard-and-fast rule.
D: We also have to consider that we make chatbots for whichever medium makes the most sense for the client. Therefore, we must propose design and interactions that complement each medium.
Q: Can you walk us through a flow you built to be completed in under 60 seconds?
D: We built a flow for setting up calendar reminders. Let’s say the user asks, “Tell me what’s on my calendar.” To present the events on his/her schedule, Florian and I must determine: What event information should be displayed? What information should be left out? In which order should we present it? Generally, less is more and we only include necessary details.
We make sure our decisions are data-driven. The different experiences and lessons we’ve learned in our roles—combined with that of others who are just as passionate about good UX—provide us with a foundation that continues to gets stronger as we work with this evolving technology.
Our chatbots team is composed of engineers, UX designers, conversation designers, NLP trainers, and project managers. We build custom bots for our clients’ unique customers. We also build for different mediums to reach customers on the channel that makes the most sense for each business. Tell us what kind of chatbot you need and we’ll take it from there.