Context and product needs
Creating a chat-style interface was not a problem, the real challenge was building user trust throught the AI.
Building trust within a chat bot by:
Including relevant suggestions for common prompts.
Displaying informative messages during loading states to explain system processes and justify any delays.
Key elements to build trust in the AI assistant include giving users control with action buttons, adding a feedback option to improve responses, and showing clear warnings with the ability to report issues to support.
Designing Data for Decision-Makers
Why charts mattered for this use case
In this project, incorporating responses in the form of charts was essential. For our finance professional user persona, who relies on quick access to reports and insights, data visualisation provided a faster, clearer way to consult large databases. By embedding charts directly into the chat, I enabled users to interpret results more efficiently and integrate the feature into their daily workflows.
Designing for clarity in complex interfaces
This was the most challenging aspect of the collapsed chat. To improve accessibility and readability, I added the ability to expand graphs to full screen, which makes complex data easier to explore and understand. This reflects my approach to work, which involves analysing edge cases of the interface to ensure the most seamless and pleasant user experience.
