Acta AI

Artificial intelligence. Dependable automation. Finance under control.

The platform does the recurring work underneath the finance function. It runs on a schedule inside your own systems, keeps a record of every run, raises what does not fit, and leaves the decisions to a person.

What it helps you get done.

Organised around the work a finance function actually owes the business each month.

Close

Closing the books

Sources pulled, recurring entries posted, accounts reconciled, and anything unexplained stopped and raised. The close becomes a review rather than a build.

Report

Building the package

Statements, the board deck, the owner or lender package, and the calculations your capital structure requires. One definition of each number, so the pages agree with each other.

Cash

Watching the cash

A rolling forecast rebuilt from live bank, receivables, payables, and payroll. What is coming, what can wait, and when a call needs making.

Operate

Running the routine

Invoicing, collections follow-up, vendor bills, payroll tie-outs, filings. Captured from the person who does them today, documented, then run on schedule with evidence filed.

Watch

Telling you what changed

Variances, exceptions, a customer that stopped paying, a contract that stopped billing. Surfaced the day it happens rather than discovered at quarter end.

Record

Keeping the evidence

Every run leaves a record of what ran, on what data, what it produced, and who approved it. The audit trail is a by-product of the work rather than a reconstruction.

Needs input before publication

Confirmation of which capabilities are generally available today versus built for a specific engagement, and any limits that should be stated. The list above reflects work running in production, but the site should not imply every item ships to every customer on day one.

People, intelligence, and automation.

Three different things do three different jobs. Confusing them is how automated finance goes wrong.

How work moves through the platform Source systems feed a nightly load. Rules and checks run automatically. Anything that does not fit is routed to a person who decides. The approved result is assembled and filed with its evidence. Your systems ledger, payroll, bank Data lands nightly, in your tenant Rules run post, reconcile, check Result assembled package, statements Filed with its evidence A person decides reviews the exception, takes the position, approves does not fit approved Automation. Runs on a schedule, the same way every time. Intelligence. Decides what needs a human. A person. Judgment and sign-off.
Nothing is approved by the platform. It prepares, checks, and escalates. A named person signs.

Control and visibility.

The controls are what make automation safe to rely on. They are not an add-on.

01

Confirmed before automated

Every process is captured from the person who performs it, written down, and confirmed by them before anything runs unattended.

02

A record of every run

What ran, when, on which data, what it produced, and who approved it. Kept as the work happens rather than assembled afterwards.

03

Exceptions reach people

Anything outside the rule stops and is routed to a named person. Silence is never treated as success.

Where it runs, and what you keep.

Inside your own systems

The platform deploys into your own cloud environment. Your data stays in your tenant. What gets built during the engagement remains in place afterwards, running for your team.

Needs input before publication

One approved commercial and technical definition, covering: what the customer owns outright and what is licensed, system and environment requirements, who administers what, data handling commitments, and what happens to the deployment if the relationship ends. This is the single most asked question by a technical evaluator and the page cannot ship without it.

After the transition.

The platform keeps running when your team owns the function.

Needs input before publication

The continuing commercial relationship is not defined. Two things need settling and they conflict in the current material. First, whether this is a recurring subscription or a retainer against a platform the customer owns, because the two say different things about who holds the asset. Second, what customer-success work is included versus chargeable, and how new automation requests are handled. Until both are settled this section should state only that the platform continues to run and that support is available.

See the full journey

Start with Discovery.

Discovery looks at the systems you run on and the work your finance function performs today. It establishes what a deployment would involve before anything is proposed.