Designing the Human Side of Autonomous Systems.

Joined AiDA as a contract UX Designer. Redesigned their product-content platform end to end - 40+ screens covering issue review, rule building, and quality scoring across a retailer's full catalog.

AiDA’s AI assisted automated system finds content errors across a retailer's entire catalog. Volume is the easy part. The hard part is the review queue - at that scale, any friction means operators stop reading, and zero friction means they stop thinking.

• Scoring PDP’s.moreA single 84% was helpful, but ability to go deeper was vital. For our platform we decomposed each product listing into description, visual, title and attributes, put it against a market benchmark, and made every sub-score clickable down to the guidance. That way a user could have as much control as they desired.
• Giving the rules back to the user.moreI wanted the query language to be something a merchandiser could operate at the individual rule level - that way they could customize by category or even deeper.
• The core screen.moreEvery suggestion is shown as three stacked parts: the rule that fired, the proposed fix, and a preview of that fix rendered inside the live product page. Accept, undo, or customize.

40+ screens across four product areas - issue triage, rule authoring, quality scoring, and team assignment. One shared shell, one nav model, one component library. Every machine suggestion traceable to the rule that produced it.