Binding authority is personal, and it stays that way. Underwriters should never let an agent bind cover beyond delegated authority, decline a risk on its own, price individuals' life or health cover without review, treat a material declaration as verified, vary terms mid-term, or use protected characteristics in its reasoning. Each commits the insurer or affects a person unfairly.
Why are underwriting red lines distinctive?
Underwriting authority is delegated personally: a named underwriter may bind risks up to set limits, in set classes. That structure exists because a wrong acceptance can cost far more than the premium, and because insurers must show who decided what.
People are also affected directly. A declined or overpriced policy can leave someone uninsured. Under the EU AI Act, risk assessment and pricing for life and health insurance of natural persons are high-risk uses, which signals how seriously these decisions are treated.
Which acts stay with the underwriter?
| Never let the agent | Why |
|---|---|
| Bind cover outside the delegated authority it has earned | Authority is personal and capped; the insurer must know who bound the risk |
| Decline a risk without an underwriter | Declines shape the book and the broker relationship, and affect applicants |
| Price individuals' life or health cover without review | High-risk under the EU AI Act; fairness and accuracy need a person engaged |
| Treat a material declaration as verified | The applicant's word is declared; material facts may need signed evidence |
| Vary terms, exclusions or premium mid-term | Mid-term changes alter the contract and need authority |
| Use protected characteristics or close proxies in its reasoning | Unfair discrimination creates legal and reputational harm |
Operations managers and collections teams protect adjacent ground; see what operations managers should never let an agent do and what collections agents should never let an agent do.
How are these lines enforced?
Binding limits sit in the insurer's own limits, so an attempt to bind above them is held for a named underwriter. Declines and individuals' life and health pricing are held at every depth. The sources bound stops a decision that needs signed evidence from resting on a declaration; the verdict names sources as the bound that decided. Mid-term changes are outside the agent's role scope. Every verdict is filed in an append-only record, which matters when a broker or a regulator asks how a decision was reached.
What does a red line look like on a real submission?
Take a commercial insurer in the Netherlands writing liability cover for small manufacturers. The agent holds depth 3 for renewals of unchanged small risks within appetite. At renewal, a client's broker notes in an email that the client has "added a small product line". The agent reads the email, which is external, and the renewal questionnaire, which says "no material change", declared.
The renewal would fall within the agent's depth if nothing had changed, but the broker's note signals a change. The agent holds the renewal rather than issuing it, and drafts a request for details. The underwriter learns the new product line is child car seats, a very different liability exposure. The renewal is re-underwritten with new terms. Without the hold, the insurer would have renewed a risk it had never assessed. French insurers apply this under the same EU framework with ACPR oversight; see underwriter training in insurance in France and in the Netherlands.
How do underwriters keep these lines in place?
Keep delegated authority current and visible in the insurer's own limits. Review a sample of the agent's depth 3 renewals each month, looking for changes it missed. Check its reasoning for proxies of protected characteristics, such as postcode in some contexts. And make sure every underwriter can explain each red line to a broker. The argument for underwriters as the deciding check is in why underwriters are the real control, and the cross-role view is in AI agent skills by role.
Where the Academy fits
Underwriters who approve the agent's work can certify through the Operator track at Pryme Intelligence Academy, with module tests set in insurance and their own role. A practice an insurer can adopt is to tie each underwriter's approval rights for the agent's depth 3 and depth 4 work to a current certificate, checked at source through VerifyDoc.ai.
Questions people ask
Can the agent renew policies automatically?
For small, unchanged risks within appetite, at depth 3 once certified. Any sign of change should hold the renewal for an underwriter.
Can the agent explain a decline to a broker?
It can draft the explanation from the underwriter's recorded reasons. The underwriter confirms it before it is sent.
How do we check for unfair proxies?
Review which factors the agent's drafts rely on and test outcomes across groups. Remove sources that act as proxies for protected characteristics.