Loading
Loading
Tune how Bob sounds, how far it can act on its own, and the guardrails it can never cross. Changes apply to your firm's Bob only, and everything Bob does stays logged and reviewable.
Hi Dana, I'm Bob, Northwind Insurance's AI assistant. Happy to help with your cover. Want me to pull up a quick quote?
From your edits, Bob has picked up that you open warmer, keep sentences short, and sign off yourself.
You open with a warm greeting, not a formal one.
You keep it plain and human, never stiff.
You sign off yourself, so the client always knows a person is there.
This is about matching your voice, not your advice. Bob never learns to change a number, a recommendation, or a compliance line from an edit, and you can turn this off any time.
Set how much Bob does on its own for each kind of work. Bob escalates anything under the confidence threshold to a person, and adverse decisions always go to a licensed adviser no matter the setting.
Bob acts on the clear cases and holds anything under the threshold for you.
Bob drafts, you send. Nothing goes out without you.
Bob runs it end to end and escalates only the exceptions.
Quotes stay indicative until a person signs off. Required to keep Bob's advice governed.
Declines, loadings, and exclusions are routed to a human before they reach the customer.
Health and personal details are stripped before anything is logged, traced, or sent to a model.
Bob has to point to the policy wording or rule behind each recommendation, or it escalates.
Shown to customers wherever Bob answers on your behalf.
The wordmark stays fixed. Your palette and name white-label the rest.
Every setting here is logged with who changed it and when. Guardrails marked Required are set by your licence and cannot be turned off.