How a chatbot drastically improved
our health app UX
Exercise, adequate nutrition and balanced mental health are essential for every one of us. The potential of technology in this field is growing each year, and, with the age of chatbots, coaching options seem to be closer than ever for everyone. It’s a terrific opportunity to reach new audiences and to provide daily advice adapted to their personal situation and state of mind.
I led this project while working at Glance Creative, where I worked as a UI and UX designer.
The brief was to create a physical and mental health app that would help overweight users to lose weight while feeling better in their mind and body. The client wanted a subscription-based model, available on both Android and iOS.
Two of the personas created for the app
The Objectives to Tackle
With the help of a psychologist working in close collaboration with the client, we identified the following goals:
- Identify and manage toxic behaviours
- Reconnect with their body
- Break old habits and get healthier ones
These goals shaped our plan for our bigger objective: Make our users feel better about their body and themselves.
We concluded that we need to focus on the user’s mental health instead of their ideal body image (allowing them, for instance, to log their emotions, and keep track of their moods).
Therefore, a classic diet and workout app wouldn’t do the job.
(Note: from there, the brief slightly changed: It shifted from being a weight loss app to a physical and mental health app)
A Chatbot as a Personal Coach: A Multifaceted Solution
The client came to us with a fairly ‘classic’ idea of the app they wanted: a profile, a fitness dashboard and a live and learn section (with resources about well-being, relaxation, and health in general).
Knowing the objective they had in mind, and that they wanted to focus on mental health, we proposed to create a chatbot instead of static resources.
The primary reasons for the chatbot feature were numerous and very exciting:
- Support, kindness, friendliness: With the right tone of voice and adapted copy, this chat feature will allow us to get closer and more intimate with our users, which is one of the pillars of this app.
- Logbook: By keeping the full history of each conversation, this chatbot also provides the benefits of a log book, which constitutes a very powerful tool in helping people to change.
- Dynamic Coaching Content: The conversations are evolutive several levels: first, the user gets adapted conversation and media content following their pre-templated answer choices, and secondly, the whole conversation tree and its copy can be easily adapted and nourished before and post launch, at any time.
- UX/UI trends: it’s not a surprise for anybody: chatbots are the new big thing. By catching this promising trend, the app would garner more interest than a classic health app.
Despite a budget increase, the client went with our chatbot recommendation. YEEPEEE. ?
Preliminay sketches for the “Coach”
The Chatbot UX Perks: Fluid Navigation, Both Way Feedback and Self Promotion.
Now that we refocused our app on a chatbot, we had to create a UX that would make the best of it. With a chatbot our navigation flow became much more fluid, on several aspects:
With our ‘Coach’, we were able to onboard the user during their first discussion with the bot. Personal info, allergies, dietary requirements and even notification authorisations were sorted. If the user refuses to complete them, the ‘Coach’ can suggest them later.
User onboarding flow, through a “Coach” conversation
For even more fluidity, we also created ‘app shortcuts’ within the conversation, sent by the bot in due course. In this way, the users were able access any app screen when needed by tapping the shortcuts.
The second huge impact that the bot had on the UX design is the ability to get real time feedback (for both the users and us). The “User to Coach” feedback would allow us to provide carefully adapted service, by regularly checking the user’s satisfaction. On the other hand, the ‘Coach’ is also able to send feedback to users. It sends them reminders, comments their (non)activity, reward or encourage them when they make progress.
The chatbot (right) added fluidity to our user journey, with dynamic routes added when needed.
The Conversation Mechanics
For the first version we chose to focus on a question / multiple choice answers mechanics. This allowed us to make the technological challenge within our grasp and to be able to focus on the user’s overall experience.
The coach can start conversation by itself following various triggers (unlocked badges, activity, weight loss plan…). He can ask questions, provide visuals about the user’s activity, suggest some action by displaying in app shortcuts …etc. On the other hand, the user can trigger conversations at anytime by logging their ‘mood’: following the type of logged ‘mood’ and its intensity, a conversation about the user’s mental health follows.
This two-way mechanics was a great way to define simple but powerful conversational UI tools. Because the client was to create the conversation tree themselves, it was crucial that the conversation rules and triggers be defined and listed clearly.
A Challenging Handover
It was decided that the design team would design the bot and CMS interface and the user onboarding conversation, and let the client manage the conversation tree and its content.
It led to two main difficulties: The first one was that our conversation tree builder needed to be visual, for our non-programmer client, the second one was that the actual content was not available to us while we were designing the interface. Because of this particular situation there was a lot of back and forth between us and the client, and it took us some time to refine the visual language for the conversation tree, but we eventually managed to create a custom visual builder, fully adapted to the client’s needs.
Reflections: Keeping the User's Needs as a Priority is Sometimes Very Hard Work
The journey from the very start of a project to the handover is never a straight line, especially when the client, the designers and the developers work in close collaboration. In this case, the emotional impact on users was crucial for the success of the product, we therefore had to make sure to iterate and let enough time at the end of the project to test the Coach interactions, which was a big challenge, and required to often take perspective on the whole project and its development.