AI System for Preparing Insurance Renewal Files for Review

The problem

When policies, endorsements, asset records and claim records are held in different sources, staff must manually align the fields before preparing a file of changes and missing information.

The insurance specialist must compare every suggestion in the changes and missing-information file with the source record; incomplete or conflicting evidence must not replace the final decision.

If the effect of an open claim on cover is not referred for separate review, the policy could be renewed using outdated risk information; the record must not progress without a human decision.

What we set out to improve

• Time taken for the draft changes and missing-information file and its supporting evidence to reach the insurance specialist for review
• How records referred for human review because an open claim may affect cover are resolved
• Number of suggestions corrected after identifying the risk of renewal using outdated risk information

How the system works

1. 01 Retrieve policies, endorsements, asset records and claim records from authorised sources
2. 02 Verify field and document versions
3. 03 Prepare the changes and missing-information file with supporting evidence
4. 04 Check for exceptions: an open claim may affect cover
5. 05 Have the insurance specialist accept, correct or reject the suggestion
6. 06 Decision gate: insurance acceptance, cover, premium and renewal
7. 07 Record the decision and source separately

Methods we used

• A source-identified data schema for policies, endorsements, asset records and claim records
• A structure that separates deterministic rules from AI suggestions
• A screen showing the changes and missing-information file alongside its supporting evidence
• A human review queue for open claims that may affect cover
• A record identifier that prevents duplicate processing

Where people stay involved

Data and security

Policies, endorsements, asset records and claim records should be processed using only the fields required. Role-based access should be applied, and unnecessary identity, health, financial and internal organisational data should not be sent to the model. Source retrieval, the AI suggestion and the insurance specialist's decision should be recorded separately.

Who this suits

Frequently asked questions