Anylytics.AI:
Data Copilot
Industry Timeframe Role
B2B, SAAS
Summer 2024
UI/UX Designer
Overview:
As a UI/UX designer, I was tasked with building an interactive chatbot that could simplify complex data insights without overwhelming users.
​
Anylytics.AI, an upcoming startup that houses and interprets complex data scopes for businesses, needed an interactive, intuitive, and social AI chatbot feature for their platform. '
​
Their Mission...
To make data interpretation easier for both excel warriors and data newbies, as they cater to businesses in a variety of industries. The goal is to democratize data analytics in the current market!
​
​​​Project Constraints
​
With Anylytics being a startup looking to quickly roll out an MVP, I had to work quickly across software, product management, and design teams.
That meant an expedited research + design process.
Who I worked with...
Project Managers
Designers
Software Engineers
Users crave clear next steps.
Interviewees who had experience with tools like JuliusAI, and Microsoft Power BI expressed that while there were a variety of useful data analytic tools, there was a steep learning curve for beginners.
​
​
Current tools lack customization.
There seemed to be a general lack of of customization for data visualizations and interfaces, often making products fall short of specific business needs.
​​​
With a limited timeframe for research, I conducted three quick interviews to understand user pain points with current data analytics and AI tools on the market.
Here's what I found...
​
​
Digging into Design...
With this in mind, I went into the design process focusing on both users new to data analytics and those with more experience.
The challenge was to make the interface intuitive, and friendly for both potential groups.*
*Another feature I focused on in intial designs was the ability to share conversations between employees. Although not implemented into the final MVP, I really enjoyed working on it.
Design System
First Iteration
Initial designs aimed to empower users to generate graphs and tables, switch between data scopes, and share data between teams.
​
I introduced tabs to simulate a familiar user experience that would let them easily switch between data scopes.
​​​
Team Feedback
​​​
The interface felt crowded and lacked clear action paths after the copilot provided data.
​
Users need a more direct way to switch between data sets, as this was a central component of the user experience.​​​
"Users will feel lost among all the tools"
Final Design
​​
To reduce clutter, I introduced collapsible navigation and sidebars, and added feature icons (sharing, visualizations, mysql code generation) to make advanced features more accessible without overwhelming users.
I also added conversation filtering + visible scope switching, to allow users to access scopes outside of conversation tabs and easily parse through past conversations.
​
Additionally, for more hestitant users, I implemented confirmation states, to make the user experience less anxiety inducing!
​​​
Reflection
The Final Designs were implemented into Anylytics.AI's MVP and is currently being presented to stakeholders! ​
​
By emphasizing intuitiveness and guidance in this AI tool, I was able to create a user experience that reflected the variety of users who might interact with these tools on a day to day basis.
The flexible interface also allows for more advanced users to bypass the hand-holding if necessary, and hide certain tools for more error prone users.
​
Why is this important?
​
​In a market increasingly reliant on AI tools, its hard to stand out as a competitor. For Anylytics.AI, this meant putting the user first, and prioritizing an experience that businesses want to come back too.
As a designer, this project provided me an opportunity to think about data accessiblity, and how we as developers, designers, and engineers can make the technical world more friendly for all. (:
