Robot Star Rating
Rate every agency's AI assistant, one to five stars, from what citizens did next.
The job
Lets a body that oversees many service providers rate each provider's AI assistant on what citizens did after using it. Providers see their rating and evidence before publication and get corrective notices, and the ratings are checked again next cycle.
The moment
Each agency's AI assistant appears as a door in a corridor, with a star plaque based on what citizens did after using it. Five plaques light up.
The assessor opens one door and the room hears the conversation that cost it stars: "the assistant wouldn't register my complaint".
That problem tripled this quarter and dropped the agency from four stars to two.
The agency's director calls the oversight body within an hour of receiving the draft report, before anything is published.
What it does
- Send the rehearsal dossier
- Issue a corrective notice
- Update plan status
- Publish a new rubric version
- Send the index to legal review
What you see
A corridor of doors with star plaques, walked like a mystery-shopper visit that replays the conversations behind every star
What it moves
The share of agencies rated four stars or above, and each agency's contacts about the assistant failing per 1,000 assistant conversations, compared cycle to cycle.
Built for
- Product & digital
- Regulators & auditors
- CX leaders & VoC
