AI System for Comparing Policies with Quotations and Applications

The problem

When quotations, applications, policies and endorsements are held in different sources, staff must manually align the fields before preparing a list of field and cover discrepancies.

The policy operations specialist must compare every suggestion in the field and cover discrepancy list with the source record; incomplete or conflicting evidence must not replace the final decision.

If an approved endorsement that reached the main record late is not referred for separate review, the wrong cover could take effect; the record must not progress without a human decision.

What we set out to improve

• Time taken for the draft field and cover discrepancy list and its supporting evidence to reach the policy operations specialist for review
• How records referred for human review because an approved endorsement reached the main record late are resolved
• Number of suggestions corrected after identifying the risk of the wrong cover taking effect

How the system works

1. 01 Retrieve the quotation, application, policy and endorsement from authorised sources
2. 02 Verify field and document versions
3. 03 Prepare the field and cover discrepancy list with supporting evidence
4. 04 Check for exceptions: an approved endorsement reached the main record late
5. 05 Have the policy operations specialist accept, correct or reject the suggestion
6. 06 Decision gate: insurance acceptance, cover and premium
7. 07 Record the decision and source separately

Methods we used

• A source-identified data schema for quotations, applications, policies and endorsements
• A structure that separates deterministic rules from AI suggestions
• A screen showing the field and cover discrepancy list alongside its supporting evidence
• A human review queue for approved endorsements that reached the main record late
• A record identifier that prevents duplicate processing

Where people stay involved

Data and security

Quotations, applications, policies and endorsements 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 policy operations specialist's decision should be recorded separately.

Who this suits

Frequently asked questions