Waston Assistant Templates
project overview
This project introduced a new feature - a catalog of templates that provides pre-built conversations covering a variety of industries, topics, and needs.
Duration
~ 8 months
Role
Primary designer, UX designer
UX designer, Visual designer, Product manager, Developer, Researcher
Collaborators
Watson Assistant is an AI chatbot software designed to make it easy for business users to build without technical expertise.
But if that’s the case, why are users dropping off after their first session?
First time user state
the problems
1
Users are intimidated by the prospect of starting from scratch.
2
Users are unaware of many features and best practices when it comes to building their conversations.
3
They’re forced to invest a significant amount of time and energy into their chatbots.
4
Users are unable to see the value of Watson Assistant in a timely manner, and abandon their projects.
Based on competitive research done by the team, nine competitors all have the same feature that we lack.
Catalogs of pre-built content to help the user get started.
the hypothesis
A catalog of pre-built content…
1
Helps users to learn and retain core concepts of Watson Assistant.
2
Reduces drop-offs as users begin to understand how this brings value to their end user.
3
Assists users in building out a robust assistant, faster.
usability testing
To validate our initial research and hypothesis, we brought our wireframes to a moderated usability test.*
Key insights
-Participants felt they had a better understanding of how to create conversations after interacting with the templates.
-Participants said they would definitely use templates in the future
-Some found the catalog overwhelming or hard to navigate
-Ability to personalize templates was important
-The previews could be more indicative of template content
*Note: This is the phase in which I joined this effort. Content before this point summarizes work done by my team, and is provided for context.
catalog design
Shortly after the conclusion of testing, the lead designer left the team and I became the solo designer in charge of this work.
Taking into consideration our testing, team feedback, and familiar catalog patterns, I began making updates to the designs.
Featured updates
Reduced visual overwhelm
Tiled catalog format
Side panel for filtering
Before
(Designed by teammate)
After
testing, round 2
For our second round of testing, I worked with one of our team’s designated researchers. With the updated designs and two Figma prototypes, we prepared for an unmoderated usability test.
The goal was to get more detail on the insights from round one, as well as see if we improved on the usability issues.
Key insights
-Significant chance that interacting with templates will improve user performance when creating from scratch.
-Some found the catalog unnattractive
-There was some confusion with navigation between templates
-No blocker issues
ready to deploy
With no blocker issues, we were ready to clean up the designs and make the final updates for our MVP.
Featured updates
Visual enhancements
Updated design to reduce use of modals (therefore reducing navigation confusion)
post-mvp workshop
After our release, we promptly started working on some design and feature updates.
The team was having trouble getting aligned on how to present one of the more technical features, so I ran a mini workshop with the team to ideate and sketch.
Agenda
Context on the issue
Inspiration hunting
Crazy 8s sketching exercise
Results
Stakeholder alignment
Idea validation
Improved experience
more updates
Our post-MVP designs began with lots and lots (and lots) of iteration. Fortunately, another UX designer and a visual designer joined to collaborate on this work.
Two out of many iterations we did, trying to incorporate the new features into the layout
the final look
Knowing that this would be the last update we could make for a while, we decided to go big or go home.
My collaborators and I were able to push through a visually exciting experience that still accounts for our users’ needs.
Ability to select multiple templates
Visual enhancements
Featured categories with correlating tags
Featured updates
Psst… I designed that green illustration!
impact
-75% of accounts that went live, used a template
-68% of all new users added a template in their first session
-250% more likely that users will reach a testing milestone
First time user state without templates
First time user state with templates
Takeaways
Adapt, but advocate
Introducing a new feature is a big deal, and everyone wants to share their two cents. I strived to be amenable to new ideas without making changes that didn’t benefit the experience.
Don’t delay value
Users were dropping off because they weren’t willing to invest time and energy into a product they didn’t see the value in. By accelerating their build, we accelerate their time to value!
Go big or go home
As I mentioned, we made some big changes to the look and feel of the catalog. What I didn’t say is that I was pretty hesitant to share something so drastic with the stakeholders. The other designers on my team taught me that the worst someone could say is no, and that’s worth it if it means advocating for a better experience.