Case study · Agentic AI · Bank, Saudi Arabia

Auto-collection automation for a Saudi bank

Collection windows, re-ageing, reversals and regulator holds turned from manual rules into automated checks.

14collection scenarios written up and automated as rules
The problem

What the client needed

Loan collections ran on rules that people applied by hand: collection windows, non-salary credits, card controls, balloon payments, reversals and regulator holds.

The result

What runs now

Each requirement carries an ID, so the bank can see which rule is automated and which still needs a person.

Core banking integrationrules enginetranscript-to-minutes tooling
What we built

The system, step by step

1

A business requirements document covering 14 collection scenarios, written from the bank's own review meetings

2

Collection-window automation first, followed by re-ageing, incident recovery, prevention of double collection, and reporting

3

SAMA rules for regulatory holds and SIMAH rules for re-ageing encoded as checks, rather than left to memory

4

Minutes and decisions produced from call transcripts, so every rule change can be traced to the meeting that agreed it

Regulations and standards
SAMASIMAH

Have a problem you want measured?

Tell us what you want to count, check or protect. An engineer will reply, not a sales script.