eazle: enabling AI to accelerate design and decision‑making

I built a white‑label interactive version of the customer app using v0 and Cursor so we can test safely with external pools via Lyssna. I coached the team to use AI tools to prototype behaviors and interactions that static Figma cannot express, and I created a custom GPT aligned to our tone and brand to act as a UX writing assistant since we do not have a dedicated copywriter. Every request starts with the problem, target behavior, and success metric; AI outputs are reviewed in crit and attached to PRDs or Confluence for traceability.

CX Research is developing four calibrated AI personas from real inputs like Google Analytics, order data, prior A/B tests, qualitative research, and surveys. We run short synthetic sessions to surface likely failure modes within hours, adjust copy or design, and decide if a small live test is warranted. Findings are shared in MS Teams and short slide readouts, and reusable templates and a safe repository are being set up so others can start from a known baseline.

This track is in build and internal testing. Adoption is early, but it already provides faster directional signal, tighter PRDs, and fewer stalls when access is constrained. One‑way‑door or high‑risk changes still require live research. Risks are accuracy and process; we mitigate by cross‑checking against prior real studies and keeping simple guardrails. Metrics remain internal and are reviewed alongside experiment readouts.

This is not about replacing research. It is about enabling the team with practical AI to learn faster, write better UX copy, and make smaller, safer bets—then validating with real users before we scale.