Automation Draft Night
Draft the intents your AI agent takes next; a receipt settles every pick.
The job
Decides on evidence which customer requests the AI agent takes on next, prices each one before it starts, and keeps or drops it after a measured first week, with a receipt the CFO trusts.
The moment
Like a sports draft, the team picks which customer requests the AI agent takes on next. Each pick is priced before it starts and gets a one-week trial.
On day seven, the first pick's card flips: for "where is my order", the bot promised to resolve 65% and delivered 71%.
Its quality score was 88 against 84 for human agents, and 6% of its customers came back, against 9% for humans.
The receipt shows $84,000 a month quoted and $71,000 delivered, after 212 resolutions were cancelled because the customer came back. The CX lead picks the next round on the spot.
What it does
- Draft an intent with a pre-registered promise and quote
- Send the agent brief and test cases
- Request the pilot
- Unlock a red (risk) intent
- Cut and return to humans
- Print the receipt
- Create or route the AI agent directly from OnClarity
What you see
Sports draft board with trading cards, a day-7 card flip, and a till receipt with voided lines
What it moves
The confirmed net monthly saving in dollars compared with the quote, and the share of contacts the bot handles with a quality score of 85 or more, not counting any where the customer came back within 72 hours.
Built for
- Contact-centre operations
- Frontline staff
- Finance
- Compliance, complaints & legal
