Bot Told Me
Catch customers quoting your bot's promise, price the exposure, fix it before precedent.
The job
Finds customers who repeat, on another channel, what an AI agent told them, checks it against what the bot actually said, and rules on it. Prices the exposure, fixes the policy gap and watches the repeats stop.
The moment
A customer posts on social media: "your bot promised me a full refund". The app finds the bot's exact words making that promise and shows them side by side.
It also finds 37 other customers who were told the same thing this week, a potential cost of $412,000.
The policy gap is fixed. Three weeks later, no new customers are repeating the claim.
What it does
- Rule on a claim
- Add a guardrail test case
- Honour or correct with affected customers
- Send the claims register to the regulator or board
- Drill the evidence
What you see
Echo map: each claim at the centre, echoes rippling out by channel; a split-screen claim court; an exposure meter in dollars
What it moves
The open dollar exposure from wrong bot promises and the liability avoided on fixed claims, days from the first repeat to the fix, and new repeats of each claim after the fix.
Built for
- Compliance, complaints & legal
- Finance
