AI System for Checking Loan Files for Missing Documents

The problem

When loan applications, identity documents and proof of income are held in different sources, staff must manually align the fields before preparing a missing-document list.

The loan operations specialist must compare every suggestion in the missing-document list with the source record; incomplete or conflicting evidence must not replace the final decision.

If a self-employment document with a different structure is not referred for separate review, an application could be unfairly screened out; the record must not progress without a human decision.

What we set out to improve

• Time taken for the draft missing-document list and its supporting evidence to reach the loan operations specialist for review
• How records referred for human review because a self-employment document has a different structure are resolved
• Number of suggestions corrected after identifying the risk of unfairly screening out an application

How the system works

1. 01 Retrieve the loan application, identity document and proof of income from authorised sources
2. 02 Verify field and document versions
3. 03 Prepare the missing-document list with supporting evidence
4. 04 Check for exceptions: a self-employment document has a different structure
5. 05 Have the loan operations specialist accept, correct or reject the suggestion
6. 06 Decision gate: loan eligibility, limit, pricing and rejection
7. 07 Record the decision and source separately

Methods we used

• A source-identified data schema for loan applications, identity documents and proof of income
• A structure that separates deterministic rules from AI suggestions
• A screen showing the missing-document list alongside its supporting evidence
• A human review queue for self-employment documents with a different structure
• A record identifier that prevents duplicate processing

Where people stay involved

Data and security

Loan applications, identity documents and proof of income 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 loan operations specialist's decision should be recorded separately.

Who this suits

Frequently asked questions