Dr. Claire Lauer — Product Manager
Sakshi Deshmukh — Design Assistant
8 weeks, Product under development
KEY HIGHLIGHTS
CONTEXT
An opportunity to improve water situation in Arizona!
Current water situation
1.0
Article on Arizona Water Situation
THE PROBLEM
Does the current solution addresses the situation effectively?
Response Format was vague and brutal.
The responses generated by current version of chatbot had to be improved significantly.
Responses aren’t relatable
The responses by chatbot were irrelevant sometimes and they missed the impact.
University's Strict Guidelines
Website is under asu domain which means it had to be designed under ASU’s strict Design Guide.
RESEARCH
Let's see what our users think…
Took 10 User Interviews
It was critical to understand where the product wasn't meeting user expectations. Understanding their mental model.
2.0
Crucial Insight from Research
What are users asking for?
Insights from interviews reflected the kind of information users are willing to learn on water situation.
2.1
Type of Questions asked by various users
CHATBOT AUDIT
Where's the current solution lagging?
Audit was conducted on Chatbot considering best UX practices to gather notable usability issues.
The Format matters the most in the chatbot, Entire learnability of the content depends on how we present the responses.
3.1
Current Conversational Format/Style
Paragraph Format
The response was unreadable. It was difficult to ensure high learnability.
What Next?
Users were not able to relate with the chatbot responses. How would users act upon the situation? What should they do? Everything prompted us to think of a solution.
WHERE'S THE CHALLENGE?
DESIGN DIRECTION
Ensuring Flexibility to users!
One of the key features is the ability to show users different types of responses. (Tell me more, Sources, Next Steps)
3.0
Heuristic Evaluation of Conversational User interface
The most vital element of the chatbot is how the responses are wrapped for the users.
3.2
Newly Proposed Conversational Format/Style
✅ Ensured Learnability
✅ Sorted the Flexibility
✅ Positive Impact
IMPACT
Let's talk Impact!
How'd we achieve Learnabilty?
Chatbot was presented at an all-hands in March 2024 and was appreciated by authorities and users.
48% Users asked follow up questions on the responses and showed curiosity.
(This also ensured that the responses were actually read by them! )
The features that we introduced (Response Format) was interacted by 56 out of 80 Users.
There were 14 out of 80 users who were bored with the chatbot responses.
(We further tweaked responses to reduce this no. to 9)
KEY LEARNINGS
What'd I learn tho?
Coding and conversations with target users, stakeholders brought me insights on how to tackle problems.
Never miss Accessibility
By Integrating speech to text inputs, we increase our engagement sessions and no. of interactions on chatbot.
It's not always about Visuals!
We had to ensure learnability and awareness in Conversational UI, we achieved it by research and tons of conversations with users.
Next project:
