An agent that reads your numbers, works out what changed and why, and brings you the report that needs a decision.

A dashboard full of numbers is not the same as knowing what to do on Monday.

Every business has more figures than it reads. Sales, margins, stock, hours, the numbers pile up in reports and dashboards nobody has the time to sit with. The dashboard shows that margin slipped or that one location is down, and then it stops, because a chart can show a number move but it cannot tell you why it moved or whether it matters. So the report gets a glance, the odd number gets a shrug, and the thing worth acting on sits unread until it is a problem.

Your agent reads the numbers for you. It gathers what your systems already produce, works out what actually changed since last time, finds the likely reason behind the move, and brings you a short report of what needs a decision. What reaches you is not another dashboard, it is the two or three things worth your attention this week, with the reason attached.

A dashboard shows you the number and leaves the reading to you. Your agent reads it and tells you what moved and why.

A dashboard is a set of dials. It shows sales, margin, footfall, whatever you wired to it, updated and accurate, and it leaves every bit of the thinking to you: which move matters, which is noise, why the dial turned, and what to do about it. The reading is the hard part, and the dashboard hands all of it back.

Your agent does the reading. It looks across the figures, works out what genuinely changed against the period before, sorts the moves that matter from the ordinary wobble, and works out the likely reason. The question it answers is not "what are the numbers" but "what changed, why, and what should you do about it."

Three businesses, and the report that lands on a Monday.

A wholesale distributor. The month's gross margin is down a couple of points. The dashboard shows the dip and nothing more. Your agent reads underneath it and finds the drop sits almost entirely in one product line, where a supplier raised the cost and the sell price was never lifted to match, and that two large customers account for most of the affected volume. It reports the line, the cause and the exposure, so the fix is a price update on one line rather than a hunt through the whole book. Nobody scanning a margin figure could see that it was one line and one un-passed cost increase.

A multi-location retail franchise. One store's sales are down against the others this month. The dashboard flags the store. Your agent works out that the dip is concentrated in one category, that it lines up with a fortnight the store was short-staffed at its busiest hours, and that the other locations held. It reports which store, which category and the likely reason, so the owner looks at rosters rather than at the whole store. A head-office figure showing "store four is down" could not tell you it was one category and a staffing gap.

A subscription business. Churn ticked up this month. The dashboard shows the percentage. Your agent works out that the extra cancellations cluster in the cohort that joined on a promotion six months ago, and that their usage had been sliding for weeks before they left, so the same early-warning signal is now flashing on a newer cohort. It reports the at-risk group while there is still time to act on it. A churn percentage on a dial cannot tell you which customers, or that the next wave is already visible.

The numbers already exist. They just sit in separate systems that do not talk to each other.

Your business generates the figures a good report needs. The catch is that they live in different places, the sales in one system, the costs in another, the stock and the hours somewhere else, and pulling them together into one picture is a chore that has to be redone every period, so it usually is not, or it is done once in a rush and half-checked.

Your agent gathers them for you. It pulls the figures from the systems you already run, lines them up into one consistent picture, and does it the same way every period, so the report is built on the same basis each time rather than on whatever someone had time to assemble. You are not asked to move your data or to key anything in; it reads from where the numbers already are.

It does this on whatever rhythm you want a report, weekly, monthly, or the morning after a period closes, so the read is waiting for you rather than being a job you have to start.

Most of what moves is noise. The job is to find the move that is not.

Numbers wobble. Sales are up a little, a cost is down a little, footfall drifts, and almost none of it means anything. Buried in the ordinary movement is the occasional change that actually matters, and the work of a report is to tell those apart, which is precisely the work a dashboard leaves to a tired person on a Monday.

Your agent does the sorting. It compares this period against the last and against the pattern of the periods before, works out which moves are within the normal swing and which are genuinely out of the ordinary, and sets the noise aside. What it surfaces is the handful of changes worth a second look, not the full page of figures that mostly say the same as last time.

By the time it writes anything up, it has already separated the two or three things that changed from the dozens that only appear to have.

Knowing a number moved is half a fact. The other half is why, and that is the half you can act on.

A figure that has moved is a question, not an answer. Margin is down: because of what? Sales at one site slipped: driven by which part of the business? The move on its own tells you to worry; only the reason tells you what to do, and finding the reason means reading across the other figures to see what lines up with it.

Your agent looks for the reason. When something has genuinely changed, it reads across the rest of the data for what moved with it, a cost that rose, a category that fell, a customer group that thinned, a period that lines up with the change, and it puts forward the likely explanation together with the numbers that point to it. Where the data cannot settle the cause, it says so and shows you the candidates rather than inventing one.

What you get is not just that something moved, but the most likely reason it moved and the figures behind that read, so you can act on a cause instead of chasing a symptom.

The report arrives short, in plain words, saying what moved, why, and what is worth your attention.

The last thing anyone needs is another long report. What the agent brings you is short: the two or three things that actually changed this period, the likely reason for each, and what appears to need a decision, in plain sentences rather than a wall of charts. The full figures sit behind it if you want them, but the point of the report is what you should look at, not everything that could be looked at.

Your agent writes it the way a sharp analyst would brief you, leading with what matters and keeping the rest to hand. You read it, and where something needs acting on, you act, or you ask it to look closer at one thread. Where you want the report to reach others, a manager, a partner, it can send it on the rhythm you set. What it never does is dress up noise as a finding or bury the one thing that matters in a page of things that do not.

A report you actually read, that tells you the two things worth knowing this week, is worth more than a dashboard you have stopped opening.

Gathers the figures · works out what changed · finds the likely reason · brings you what needs deciding

It reads the numbers so you do not have to, and stops where the decision is yours.

• It gathers the figures from the systems you already run, sales, costs, stock, hours, whatever the report needs, without asking you to move your data.
• It builds the report the same way every period, so each read sits on the same basis as the last and the comparison is honest.
• It works out what actually changed against the last period and the pattern before it, and sorts the moves that matter from the ordinary wobble.
• It finds the likely reason by reading across the rest of the data for what moved with the change, rather than reporting the number alone.
• It says when the data cannot settle a cause, and shows you the candidates instead of inventing one.
• It writes it short and in plain words, leading with what needs a decision and keeping the full figures to hand behind it.
• It flags what appears to need action, an at-risk group, an un-passed cost, a location drifting, while there is still time to act.
• It sends the report on the rhythm you set, to you or to whoever you choose, weekly, monthly or after a period closes.
• It looks closer on request, taking one thread from the report and digging into it when you ask.

Your numbers, your periods, what counts as normal, and what you want watched

What a useful report looks like is particular to you. The figures you steer by, the periods you compare, how big a move counts as worth flagging, the parts of the business you want watched closely, and the things you would rather see the moment they move than at month end.

Your agent is built against those. It reports on what you actually steer by, flags what you would want flagged, and when the numbers you care about or the way you read them change, it changes with you.

Every report shows its working, and you decide what it watches and when it speaks

Each report shows where its figures came from and how it reached each read, so you can check the reasoning behind any finding rather than take it on trust. Where a conclusion rests on an assumption, it says so.

You set what it watches and when it speaks. It can report on a fixed rhythm, or watch certain numbers and raise something the moment it moves, or stay to the summary and dig deeper only when you ask. Where it is unsure of a cause, it does not assert one. It shows you what it found and what it could not resolve, so you are never handed a confident answer the numbers do not support.

Who this is for

This suits any business that produces more numbers than anyone reads, where the figure that matters hides in a report nobody has time to sit with. Distributors and wholesalers watching margins across many lines and customers. Multi-location operators comparing sites and wondering which drift is real. Subscription and membership businesses reading churn and usage for the signal before the cancellation. Anyone whose best analyst is the one who looks at the numbers and says the two things worth knowing, rather than printing all of them.

If your reports get a glance and the thing that mattered was in them, the read is the thing to build first.