RIBO's May 2025 guidance on responsible AI use asks for three things: a customer should know when they are engaging with AI rather than a person; a licensed broker should stay in the loop; and anything AI-generated should be overseen by a licensee before it reaches a client.
The system is built to sit inside those lines rather than near them.
- It identifies itself. Every message says who is writing and on whose behalf.
- It never quotes a premium, recommends coverage, or gives advice. Those are licensed acts and they stay with your licensed people.
- It escalates rather than guesses. Any coverage question, any complaint, any claim, anything the classifier isn't confident about — the sequence halts and a person is notified.
- Take-All-Comers is untouched. Nothing here selects markets or produces a rate.