During a 2-month internship, I designed the vision prototype for the desktop version of an AI-powered simulation platform for wealth advisors. Expanding from an existing mobile app, the goal was to scale voice-enabled, scenario-driven coaching and deep contextual feedback into a desktop experience.
Jump to the solution →Process
The project involved active global collaboration, wherein I got design mentorship from the India and NYC teams while aligning with the NYC-based product team to define scope, vision, and my desktop iterations.
*Adopting new ways of working in the process of it all
Research
To ensure platform parity, we needed to adapt the mobile voice simulation into a desktop side-panel. The challenge was designing a vision prototype that eases first-time advisors into AI simulations while establishing intuitive gamification and feedback loops.
Before designing the desktop side-panel, I analysed both the existing Minimum Viable Product and the early future vision concept on mobile to pinpoint core gaps.
*Due to company policy reasons, I can't show you the audit, but I've mapped the primary app flow here
Unexplained app-specific jargon and an ambiguous landing page create high friction for first-time users.
Badges and leaderboards were visually present but lacked underlying logic, point triggers, or clear progression rules.
No real-time processing cues left users unsure if the AI was listening during the session.
Scenario browsing lacked clear learning curves or game tiers, making practice feel like a chore rather than skill-building.
Referencing
Leveraging the existing design system components, I explored layout and interaction variations to test different structural ideas on both Figma and Github CoPilot.
For this, I studied patterns within legacy products, AI voice chat flows across platforms like Claude and Hume AI as well as gamification elements within apps like Duolingo.
*Referencing exercises including flows and static elements
Solution
Recreations of the final flow created. In my hand-off files, I'd provided screens, a component set with states, hierarchy and scoring logic wherever applicable along with annotations.
Closing
Run formal testing with new and seasoned wealth advisors to validate whether the new version improves their practice experience.
Due to current regulations, a voice-to-text model was used — voice-to-voice would create a more real-world feel.
Allow users to create custom practice scenarios imitating a client's risk appetite, demeanour, and portfolio.
Contextual pointers before a session starts, based on the advisor's previous performance.
In my short duration at the internship I got to learn a lot about the ways of working at the organisation, including new tooling and collaboration requirements. These were some of my most exciting learning takeaways.
Incorporated VS Code and GitHub Copilot alongside Figma to test interaction logic and validate edge cases early.
Refined how I pitch ideas, document rationale, and collaborate across India and NYC time zones.
Learned to apply an existing design system while building gamification patterns from scratch.