The Pilot · $7,500 · one close cycle

Pick the process that costs you the most. We run it next month.

One named process, run against your books on your close calendar, delivered in the month you would have delivered it anyway. Not a review of it. Not a plan for it. The work itself, done once, so you can see what it looks like before you decide anything else.

Book a Close Teardown See what has been built

The pilot is scoped in the teardown, not on this page. Forty-five minutes, free, and you leave with a written page about your close whether or not you buy anything.

What running it means

Four things that are true before the first entry is proposed.

Read access, and one write pathpropose → a named person approves → post Nothing reaches your ledger without a person with the right role approving it first. There is no second path, and an action the system does not recognise fails closed to the most gated tier rather than guessing.

The question “what did it change without asking” has a one-word answer.

Your calendar is the calendaryour close dates → our schedule → your deadline The pilot runs against the close you already have, on the dates it already falls on. Nothing about your cadence changes to accommodate it.

You are not running a project alongside the close. You are running the close.

Your rules, not oursyour materiality → your approvers → your review steps Preparer and approver stay separate, with the review policy set per entity. Your thresholds, your sign‑off, your chart of accounts, on the ledger you already run — QuickBooks Online, Xero, NetSuite or Sage Intacct. Nothing is ripped out and replaced.

Nobody has to learn a new way of being right.

Everything is written downthe run → each proposal → the approval → the sync Every proposal carries before and after, its source data, the role required to approve it, a content fingerprint and its execution record — recorded as one chain you can walk backwards.

When the audit committee asks what the AI touched and who approved it, the answer is a document, not a recollection.

Pick one

The processes that make good first pilots.

Each of these already exists as a working module. That is the point of choosing from this list rather than describing something new — a first pilot should be a thing that has been built, not a thing that has been promised.

Flux and variance commentarycompute the movement → find the drivers → draft the commentary → you edit and sign Period-over-period movement computed from the ledger, decomposed into the accounts and transactions that caused it, with the narrative drafted against the actual drivers rather than assembled from last month’s file. Claims that cannot be verified against the underlying facts are marked as unverified.

The commentary starts from the numbers instead of being reconciled to them afterwards.

The thirteen-week cash forecastopen AR and AP → place by entity DSO and DPO → weekly grid → override a cell Open invoices and bills placed into weekly columns using each entity’s own collection and payment behaviour, with a pattern-based fallback for payroll and rent, and saved overrides where you know something the ledger does not.

The weekly cash question is answered from the ledger rather than reassembled every Monday.

Intercompany and the entities that will not tiematch across entities → surface the breaks → propose the entries → approve Matching run across entities that do not share a chart of accounts, with the unmatched population surfaced as a worklist rather than a difference on a summary line.

The eliminations stop being the reason the close has a bad Thursday.

The board and sponsor packassemble from the closed period → drill to the transaction → issue Reporting assembled from the period you just closed, with every figure traceable back to the transactions underneath it, rather than copied into a deck and re-checked by hand.

The pack and the books cannot disagree, because they are the same numbers.

The covenant certificatedefine the formula over your COA → compute live → pass, warn or breach Built with a formula builder over your actual chart of accounts and computed from the ledger between reporting dates, with a standard senior-loan template as a starting point rather than an inference.

You watch headroom tighten in week two, not on the certificate.

If the process costing you the most is not on this list, say so in the teardown. Sometimes the answer is that it should be the second pilot rather than the first, and sometimes it is that we are the wrong people for it. Both are useful answers to get in forty-five minutes.

At the end of the month

Three things, and all of them are yours.

The output, delivered on the date it was due

The thing itself — the commentary, the forecast, the pack, the certificate — produced on your close calendar and reviewed by your people before it goes anywhere.

Why it is firstIt is the only part that answers the question you actually have, which is whether this works on your books.

The evidence pack behind it

What ran, what it proposed, who approved it, what changed, and what it declined to touch. One chain, in the order it happened.

Why it is secondIt is the workpaper your auditor asks for and the one most finance teams cannot currently produce for anything automated.

A written read on the next two

Which processes are worth doing next, in what order, and what each is worth — written against what we saw in your books, not against a category.

Why it is thirdIf the honest answer is that there is not a second one worth doing, that is what it will say.

All three are yours whether or not anything follows. There is nothing to renew and nothing to cancel — a pilot is one month, and it ends.

Honest limits

When this is the wrong thing to buy.

Four situations where we will tell you so in the teardown rather than take the engagement.

The books do not close on a cadence yet A pilot runs against a close. If the close is still being assembled differently every month, there is nothing stable to run against and the pilot measures the wrong thing.

Fix the cadence first. That is not work we need to be paid for.

You want a strategy document This is not an assessment, a readiness review or a roadmap. It produces one month of finished work and a short written read. If what you need is a document that surveys the whole function, buy that instead — from someone else.

You would be paying us to write, and writing is not the expensive part.

The process changes shape every month Some work is genuinely bespoke each period. It can still be automated eventually, but not first, and not inside one cycle.

Pick the boring one. The boring one is where the hours are.

There is a funded ERP migration on the calendar If the ledger underneath is scheduled to be replaced, wait for it. Building against a system you are leaving is a way of paying twice.

Come back after the migration. The offer will still be here.

It starts with forty-five minutes and costs nothing.

Bring your close — the checklist, the calendar, the entity you argue with most. We go through it together and you leave with one written page about it, whether or not there is a pilot at the end.

Book a Close Teardown See the whole ladder