Your accounts say what happened. AI can tell you why.

Your monthly management accounts tell you what happened: sales, costs, profit. They rarely tell you why, or what happens next. Why profit went up while the bank balance went down. Where you'll finish the year, and what it takes to hit budget. Which number is about to go wrong. What a price rise would be worth. AI can answer those questions in plain English, every month, from the accounts you already get, plus a handful of numbers about how your business works.

Aigentic is a UK AI advisory that builds monthly board packs for owner-run businesses. Below is the kind of thing a pack finds, drawn from our client work with an owner-run UK business with three divisions. We've changed the figures and details so the business can't be identified. The pattern and the questions are real.

"We made a profit. So why is there less cash?"

Over eight months the business made about £420,000 of profit. Over the same eight months its bank balance fell from about £510,000 to £230,000.

The accounts showed both numbers. Nothing in them connected the two. The main reason was money owed by customers: it had nearly trebled, from about £700,000 to £1.9 million. The work was done and invoiced, and the profit was real. The cash just hadn't arrived.

The accountants' pack had no cash flow page, so nobody had seen it happen month by month. The fix we recommended was simple: a monthly cash view in the board pack, built from the balance sheet the accountants already produce.

"Where will we finish the year, and what does it take to hit budget?"

The accounts covered eight months. The board wanted to know about twelve. So the pack carried the year forward: what's already won, what's in the pipeline, and what the business usually wins in the last four months.

Full year	Budget	Forecast	Gap
Sales	£7.8m	£7.4m	(£0.4m)
Profit	£700k	£610k	(£90k)

On its current course, the business would miss its profit budget by about £90,000. That's useful to know in September. It's useless in January.

Then the part that turns a forecast into a plan: what closes the gap. At the business's average gross margin of about 35%, £90,000 of profit means roughly £260,000 of extra sales. That's about three more contracts of average size, signed by the end of October so they can be delivered before year-end. Or a smaller number of contracts plus the price rise described below. Which mix is the board's call. The pack's job is to show the options and then track whichever one they choose, every month.

"What's about to go wrong?"

This is the early warning, and it comes from the numbers that move before profit does.

Sales and profit are the last things to change. Before them come leads, proposals sent and contracts won. So the pack tracks those every month against plan, and flags any that start to drift.

For example: proposals sent drop from around 14 a month to 9 over the summer. Sales don't move yet, because work takes a couple of months to turn into invoices. The accounts would show the problem in the autumn. The pack shows it in the month it happens, while there's still time to do something about it.

"Why is our smallest division behind plan?"

Three reasons, and only one of them was expected.

Gross margin was 23%, against the 33% the plan assumed.
Materials took about half of every pound of sales. That's structural for the product, and it wasn't news.
Outside equipment hire took 8% of sales, against 4% and 2% in the other divisions. That was news. It raised a question for the person running the division: what's being hired in, and could the business's own equipment do it?

Then the useful part: what each lever is worth. Because half of every pound goes on materials, price matters far more than buying. On about £1.4 million of annual sales, a 5% price rise was worth about £70,000 a year. A 5% saving on materials was worth about £35,000. A price rise of just over 5% would, on its own, take the division to break-even.

That's the difference between "this division is losing money" and "this division needs a 5% price rise".

What did it take?

Less than you'd think. The accountants' monthly pack, which the business already paid for. Ten numbers a month from the team: leads, proposals, contracts won and sales by product. And five cost figures, given once and checked every quarter.

No finance hire, and no new software for the team to fill in. The first version asked for far more every month. It rarely got filled in, so we cut it back to what the board actually needed.

Can you trust what the AI says about your numbers?

Only if it shows its working and admits what it doesn't know. Three rules we hold the pack to:

Every figure traces back to the accounts or to a number someone gave us. Nothing is invented to fill a gap.
Estimates are labelled as estimates, and checked against the real figures every quarter.
When two sources disagree, it asks. Here, the accounts showed one sales figure for a division and the team's own monthly sheet showed a much lower one. The pack didn't pick one. It put the question to the person who would know.

Try it now: a prompt for your own accounts

Paste your latest management accounts into ChatGPT or Claude with this prompt. Use a business account with training switched off, and remove anything you wouldn't want shared.

You are a finance director reviewing my company's management accounts. Be direct. Plain English, no jargon.

1. Profit versus cash: did profit and the bank balance move the same way? If not, explain why, using the balance sheet (money owed to us, money we owe, stock).
2. The year ahead: if the rest of the year looks like the year so far, where will we finish against budget? What would it take to close any gap?
3. Early warning: which number moved most against last year or the plan, and what question should I ask about it?
4. Levers: what would a 5% price rise be worth over a year, compared with a 5% cut in our biggest cost?
5. Questions: list anything in these accounts you can't explain from the figures alone. Don't guess.

It won't know your business the way a board pack built on it does. But if question 1 surprises you, you've found the gap this article describes.

Key facts

Drawn from Aigentic's board pack work with an owner-run UK business with three divisions, 2026. Figures changed to protect the client.
Profit and the bank balance moved in opposite directions because money owed by customers nearly trebled.
Carried forward to year-end, the business was heading for a profit miss of about £90,000, which meant roughly three more average contracts signed by the end of October.
Leads, proposals and contracts won move months before sales and profit, so they give the early warning.
Where materials take half of every pound of sales, a 5% price rise is worth about twice a 5% materials saving.

FAQ

Why has my profit gone up but my bank balance gone down? Usually because the profit is sitting somewhere other than the bank: in money customers owe you, in stock, or in loans repaid. In one business Aigentic worked with, the bank balance roughly halved over eight months of healthy profit because money owed by customers nearly trebled.

How do I know if we'll hit budget? Carry the year forward from what's already won, what's in the pipeline and what you usually win in the remaining months, then compare it with budget. Do it monthly, not at year-end. The useful output isn't the gap but what closes it: how many more contracts, by when, or what price change.

What are the early warning signs that we'll miss budget? The numbers that move before sales: leads, proposals sent and contracts won. Work takes weeks or months to turn into invoices, so a drop in proposals shows up in the accounts a quarter later. Tracking those monthly against plan gives you time to act.

Can AI explain my management accounts? Yes, if it can see the accounts and a few numbers about how the business works, such as sales by product and main cost rates. It should label estimates, trace every figure to a source and ask when two sources disagree, rather than guess.

Is a price rise or a cost cut worth more? It depends on your margin. When materials take half of every pound of sales, a 5% price rise is worth roughly twice a 5% materials saving, because the price rise applies to every pound and the saving only to half of it.