Owed vs Told
Weigh what passengers were owed against what bots and agents said.
The job
Catches both ways disruption conversations go wrong: rights left unsaid and rules made up, by human and AI agents alike, including another vendor's. Puts a liability figure on each, fixes them and proves the fix on next week's conversations.
The moment
A pair of scales weighs disruption conversations. One pan holds rights passengers were owed but not told about; the other holds promises that were made up.
The made-up side drops: 41 callers were told "full cash refund within three days", and no such rule exists. The exposure is $96,000.
The QA lead types the same question into the bot, which says it again, live. The fix goes to the knowledge base owner that afternoon, and the next day the bot gives the right rule.
A week later, the scales re-weigh new conversations: the made-up side is empty, and the owed side is down by half.
What it does
- Queue agent coaching
- Fire make-good callback
- Send bot knowledge fix
- Ask the bot live
- Send regulator file
- Open a pan's conversations
What you see
Balance scales with two evidence pans, a live Ask the Bot console and a monthly regulator file
What it moves
The weekly dollar exposure on each side, compensation owed but not told and made-up promises, and the rate of made-up promises per 1,000 AI conversations.
Built for
- Contact-centre operations
- Compliance, complaints & legal
- Finance
