Note: This project involved a feature that was still unreleased and under internal review when my role ended. Specific product and feature details have been withheld or generalized accordingly.

The Problem
At N-able, I was asked to help validate a new AI-driven feature within the core platform — a concept novel enough that a standard static prototype wouldn't give test users a believable sense of interacting with something that actually worked. Building real backend functionality before validating the concept would have taken far longer than the testing timeline allowed.

My Role & Constraints
I was tasked with designing this experience, and began using Claude to build a high-fidelity, fully offline prototype that gave test users the illusion of a working product — letting the team validate the concept through realistic testing without committing engineering time to something unproven.

Process & Key Decision
The prototype simulated real functionality: initiating a process, showing it actively running, and surfacing notifications when a manual action was needed — recreating the rhythm of using the finished feature without any real backend behind it. I started by shaping the overall structure in Figma, then brought that into Claude Code, using the design system's MCP and agent tooling to keep the prototype visually consistent with the team's established design language throughout. I worked alongside a colleague prototyping a related capability for a different part of the product portfolio, and we deliberately coordinated to keep interaction patterns consistent rather than solving the same problem separately.

Outcome
Although the project didn't reach formal user testing before my role ended, it was well received throughout development. Regular syncs with the product manager surfaced mostly minor refinements rather than fundamental rework, and the broader team responded positively to the direction. One moment stood out: after being blocked on missing information, I received it one day and delivered a fully updated, aligned prototype the next — a turnaround specifically praised by the stakeholder involved. That pace, made possible by prototyping directly in Claude Code, would ordinarily have taken considerably longer through the team's usual process.

Reflection
If the project had continued, the priority would have been getting the prototype in front of real users — the underlying concept was new enough to its target audience that testing needed to validate the value of the idea itself, not just its usability. This work also fed into a personal exploration of how to standardize AI-assisted prototyping more broadly across a UX practice.