SHEAF · AI CREDIT-RISK CONTROL

Sheaf PortfolioIQ

A chief risk officer checks lending policy against capital requirements at two levels. The macro view, where every customer's income and commitments are added up to show how much of the book would fall into financial stress. And the micro view, one customer's income and spending. Sheaf sits over the AI making these decisions, so a recommendation that contradicts what a customer can afford is stopped before it reaches them.

Illustrative view for a specialist lender · representative lending book, sample data · stress bands and behaviours follow the Castlight affordability model

2,204

Customers in the book

3.5%

In the red (Stressed) at current policy

£1.10m

Capital at risk (expected credit loss)

Unsafe AI credit decisions stopped (last 30 days)

88/100

How well AI decisions hold together

Macro · stress-test the book · every customer's income & commitments, added up

Apply an increase to all expenditure: loans, mortgages, groceries +0% Move the slider to model a rate rise or a cost-of-living shock, and watch customers move toward the red as their spare money disappears. Click a band to see the customers in it. Click a customer to see the income and spending behind their score.

Customers in the red

3.5% of the book · Stressed band

Capital at risk

£1.10m

expected credit loss across the book

Policy headroom

Within appetite

capital requirement covered at this shock

Customers in the Stressed band

Customer	Money in each month	Left over	Affordability score	Bureau score	Warning signs

One customer · the income & spending behind the score

Where Sheaf sits · the AI decision is checked before it goes out

Stopped action one example from this book, caught before it reached the customer

An AI credit agent recommended a £6,000 debt-consolidation top-up for M‑4471.

It was confident, and it was built on a contradiction no single step caught: two AI agents in the chain used two different incomes.

Affordability agent

Income £2,780 · approve, comfortably affordable

these never
compared notes

Open-banking read

True income £2,180 · £600 was itself a loan

Sheaf checked the two agents against each other and found they did not agree at that exact point. Sheaf stopped the approval and passed it to a human specialist under Consumer Duty (FG21/1). The customer was never sent an offer that would have made things worse. Every stopped and released decision carries the evidence of why, ready for the FCA, the board, and buyers doing due diligence.

Why a CRO runs the book on Sheaf

Macro and micro, one view

Policy you can stress-test in seconds.

The whole book added up against capital requirements, and one click into any customer's real income and expenditure. Set risk appetite from where the capital actually sits.

Measure before you lend

Affordability the score can't see.

Income replaced by loans, rising gambling, bank charges, less and less spare money. The behaviours that put a customer in the red are shown before an AI acts on a score that looks fine.

Use AI, and show it is safe

Speed with an evidence trail.

Contradictions are caught and stopped, and every decision carries its evidence. The lender can use AI across the book faster with the risk measured, and the same record answers the regulator.