Which Year Counts? How Drake Turned an AI Averaging Error Into a Completed Purchase

The Client

Our client was a self-employed management consultant in his early forties, trading as a sole trader with just over two years of accounts behind him. He had a solid deposit from savings, a clean credit history and no adverse markers.

Confident with technology, he had spent several evenings using an AI chatbot to research his property purchase before approaching a high-street lender directly. He arrived well-read, fluent in the terminology, and sure he already understood how his mortgage application would be assessed.

The Objective

He wanted to buy a home for himself at 85% loan-to-value, using his self-employed income to demonstrate affordability. His two years of accounts showed a business that was growing: a solid first year, and a materially stronger second year as his client base expanded.

His AI research had told him, correctly in the abstract, that lenders assess self-employed applicants on their accounts and that two years of trading history is often enough.

On paper, it looked straightforward.

The Complication

The chatbot had answered every question he thought to ask. The problem was the one it answered without being asked: how his income figure would be calculated.

It told him lenders would take an average of his two years, and it presented that as simply how it works. Because his first year was much weaker than his second, that average dragged his assessable income down, and on that number his affordability fell short of the 85% loan he needed.

He was heading for a decline, convinced the shortfall was a fact of his finances rather than an artefact of one lender’s method. This is precisely the trap the AI could not warn him about: it gave him a general rule and never mentioned that the rule is not universal.

How We Assessed the Options

When he came to Drake, we did the thing an AI structurally cannot: we asked the questions behind the question.

Within one conversation we established that his income was on a clear upward trajectory, that the weak first year reflected start-up costs rather than a struggling business, and, crucially, that lenders do not all average.

Some take an average of the two years; others, where income is rising, will assess on the latest year alone.

That single distinction was the difference between a decline and a comfortable approval. We then compared the realistic routes open to him.

Identifying the Right Route

Step 1 – Reapply to the original lender. Ruled out. Their policy averaged the two years as a matter of fixed criteria, so the shortfall would not resolve however the application was presented. Persisting would simply have produced a second decline and a further credit search footprint.

Step 2 – Reduce the loan-to-value. Ruled out. Lowering the borrowing to fit the averaged figure would have meant finding a materially larger deposit he did not have, or abandoning the purchase. It solved the lender’s arithmetic at the client’s expense, not the other way round.

Step 3 – Place with a lender that assesses the latest year. Recommended. A whole-of-market search identified lenders that, where self-employed income is increasing, base affordability on the most recent year rather than a two-year average. On his stronger second-year figure, the same accounts comfortably supported the 85% loan he wanted.