Perspectives from the practice.
Four articles on what we see across the engagements we run — what's actually broken in most SME financial reporting, how AI surfaces it, why AI can't be the check that signs off, and what deal-readiness actually requires. Written by Timothy Messieh, CA. New articles each quarter.
The check that can't say no
I wanted to know whether an AI could be the second set of eyes on a payment run — the reviewer who catches the bad invoice before the money leaves. So I ran the test: five reviewer configurations, three models, real payment packets with problems planted and clean ones mixed in. It caught every trap I set. Then it blocked two payments that were completely fine. The obvious lesson — “AI over-flags” — isn't the useful one. Fraud, groupthink and quiet error don't need a criminal; they need a room where nobody says no.
Read the articleWhy your cash flow forecast is always wrong (and how to fix it)
Every 13-week cash flow forecast I've seen in twenty years of advisory work has been wrong. Not slightly wrong — structurally wrong. Not because the people building them weren't careful, but because the way most forecasts are built makes them wrong by construction. Here's the mathematical structure of the problem, and what a forecast-as-function-over-live-data approach looks like instead.
Read the article90 days to deal-ready: the due diligence preparation timeline
Most founders preparing to sell their business underestimate what deal-readiness requires by a factor of three. The owner-driver who's been running the business for twenty years tends to think "the books are good, we'll be ready" — until the first buyer asks for normalised earnings and the answer is silence. Here's what a 90-day deal-prep actually looks like, why the 12-month version is worth 20-40% more on enterprise value, and how fee deferral aligns commercial interests.
Read the article