Underwriters should rehearse the incident where an agent binds business the insurer no longer wants. A 90-minute tabletop built on a missed appetite change lets the team pause the binding agent to draft-only, escalate to the underwriting manager, review every policy bound in the window, decide what to do about live contracts, and report the exposure.
What does an underwriting drill rehearse, and who owns it?
In many insurers, an underwriting agent quotes standard small risks against the rating rules at depth 2 and binds those inside the appetite guide at depth 3, within the insurer's own limits on sum insured and class. Anything outside appetite, above authority or unusual is referred to an underwriter. Underwriters decide referrals, pricing exceptions and declines.
Underwriting incidents have a feature most others lack: once a policy is bound, it is a contract. The insurer cannot simply reverse it. Cancellation, if allowed at all, follows the policy terms and the duty to treat customers fairly. So the drill is as much about deciding what to do with live contracts as about stopping new ones.
The underwriting manager owns the drill. A senior underwriter facilitates. Underwriters from the affected class, someone from the broker or distribution team, a compliance contact and someone who reports exposure to reinsurance take part. If drills are new to the team, the operations team drill is a simpler place to start, since it has no live contracts to unwind.
The scenario
Friday, 14:00. A claims handler notices a new flood claim on a small commercial property policy bound three weeks ago. The postcode sits in an area that underwriting excluded from appetite at the start of the month.
The cards reveal the rest. The appetite guide was updated with the new exclusion, but the update sat in a document the agent did not draw on; the agent's sources still held the previous version of the guide. Since the change, the agent bound 30 policies in excluded postcodes, each within its own limits on sum insured. The flood claim is the first loss.
The drill step by step
- Person (facilitator): reads the claims handler's note and starts the clock.
- Person (underwriter): escalates to the underwriting manager with the policy number.
- Person (underwriting manager): pauses the binding agent to draft-only. Agent, depth 2: quotes are still drafted, but every bind is held for an underwriter.
- Person (underwriting manager): adds a refusal for excluded postcodes to the insurer's own limits so the agent refuses those quotes outright while the source is fixed.
- Agent, depth 1: lists every policy bound since the appetite change, from the record, with the version of the appetite guide each bind relied on. Person (underwriter): checks the postcodes against the current guide.
- Person (underwriters): review each of the 30 policies and recommend: keep on risk, review at renewal, or consider cancellation where the terms allow and it is fair to the customer.
- Person (compliance contact): checks the recommendations for fair treatment before any customer or broker is told.
- Person (underwriting manager): records decisions and reports to the head of underwriting, the reinsurance contact for aggregate exposure, and the risk function.
- Person (facilitator): runs the debrief and lists the source and bound changes.
| Step | The agent does | Depth | Stays with a person |
|---|---|---|---|
| Alert | Has bound policies outside appetite | 3 | Noticing the claim |
| Pause | Drafts quotes; every bind held | Dropped to 2 | Deciding to pause |
| Interim limit | Refuses excluded postcodes | 2 | Setting the refusal |
| Scope | Lists binds and guide versions | 1 | Checking postcodes |
| Policy review | Nothing | 2 | Keep, review or cancel |
| Fairness check | Nothing | 2 | Compliance review |
| Report | Nothing | 2 | Exposure reporting |
What is distinctive about an underwriting incident?
The harm is already contracted. Thirty policyholders bought cover in good faith. Cancelling mid-term because the insurer's own guide was not loaded may be neither permitted nor fair. The drill should make underwriters argue each option with the customer in mind, not only the loss ratio.
The interim limit is the clever move. A pause to draft-only stops automatic binding but still produces quotes in excluded areas for underwriters to decline one by one. Adding a refusal for those postcodes to the insurer's own limits stops the agent drafting them at all. Underwriters should practise making that kind of narrow, fast change and filing it.
Version matters. The record shows which version of the appetite guide each bind relied on. Teams should practise reading that, because it answers the question every reviewer asks later: did the agent break a rule, or follow an old one?
Which controls does the drill test, and what evidence do they leave?
The pause. Evidence: the limit change in the record, with name and time.
Interim refusal. Evidence: the new limit, who set it, and the refused quotes that followed.
Scoping. Evidence: the list of binds with the guide version each relied on.
Policy decisions. Evidence: each underwriter's recommendation with reasons, and the compliance check.
Return to binding. Evidence: the updated sources and a fresh certificate before depth 3 returns. Underwriters who approve binds during the pause should hold a current reviewer certification, such as Pryme Intelligence Academy's Certified Pryme Operator.
What should the underwriting manager measure?
- Days between the appetite change and its detection.
- Policies bound outside appetite before the pause.
- Time from the pause to the interim refusal.
- How many of the 30 policies had a decision recorded within a week.
- Whether every appetite change now names the person who updates the agent's sources.
Where do underwriting drills go wrong?
Cancelling by reflex. The team's first instinct is to take the risks off cover. The drill should make them face fair treatment before anyone writes to a broker.
Forgetting the broker. Brokers placed many of these risks. They need a consistent message from one contact, not a call from each underwriter.
Treating the guide as a document, not a source. The fix is not "remember to update the agent". It is a rule that the appetite guide reaches the agent through one owned source, with each change filed.
Use the AI agent incident drill script template to run it, and compare the compliance team drill for the reporting side and the IT team drill for the source-loading side. Underwriters joining the review rota should work through the reviewer onboarding checklist for underwriting teams. The topic guide on training teams for AI agents shows where drills fit.
Questions people ask
Can the agent ever cancel a policy?
It can draft a cancellation for an underwriter to approve. Cancelling a live contract is hard to undo and affects a customer directly, so it stays with a person.
Should reinsurers be told about a drill result?
Not about the drill. In a real event, the aggregate exposure may matter to reinsurance, so the reporting step should name who decides that.
How often should underwriters drill?
After every significant appetite or rating change, and at least twice a year. Appetite changes are the most common source of this kind of incident.