Month end close automation that shaves days off the close.
We build AI agents that prorate the invoices crossing your close boundary, dedupe the related vendor bills, and propose AP accruals for review — so your books close days sooner.
Official services partner of the platforms defining AI
NVIDIA
Anthropic
AP accruals take two or three days of the close.
From finance teams we have sat with: CFOs and controllers at multi-branch healthcare and services companies, where the close calendar is fixed and the accrual work is not.
Two or three days on AP accruals
Every month, two or three days of the close period go to AP accruals: finding the bills that cross the boundary, splitting them across periods, proposing the entries. The calendar does not move; the work does not shrink.
The prorate is done by hand
An invoice that spans two months gets split by a person with a spreadsheet — branch by branch, line by line. Multiply that by every entity you close.
The same vendor, billed twice
Related invoices from the same supplier arrive separately and get deduped by eye at the busiest point of the month. Miss one and the accrual is wrong.
No way to check the pricing
The bills keep coming, and nobody has a way to check that every line matches the supplier price file. The overcharge surfaces long after the money left — usually as a next-month adjustment.
The entries still get keyed
After the accruals are agreed, someone keys the journal entries into the GL by hand — the last manual step of a manual process.
“We need to get our books closed quicker”
Heard from the CFO of a multi-branch care provider. The close is not slow because the team is slow — it is slow because the work waits for people who are already full.
Two or three lost days is a reporting delay, not an inconvenience.
What the manual close looks like at a multi-branch provider. Your numbers will differ — tracing one close cycle puts figures on yours before anything gets built.
From an anonymized engagement — the CFO of a multi-branch hospice and home-health provider
Accrual automation: we encode your close calendar.
No new system for your team to learn. The checks your best accountant runs at month end become the system — AI agents run them, people approve the entries.
- 01
Trace one close
We follow one month-end close: which invoices crossed the boundary, how accruals were proposed, where the two or three days actually went.
- 02
Encode the close rules
Proration logic, dedupe rules, accrual thresholds, approval routing — documented against your close calendar and your GL, encoded as rules you own.
- 03
AI agents draft the accruals
Each boundary invoice is prorated, related vendor bills deduped, and accruals proposed for staff review. Approved entries batch-upload into Sage Intacct or your GL.
Balance sheet reconciliation software rents you the checklist. Your close rules are the product.
Close software ships checklists, task trackers, and a dashboard of what is late — licensed per seat, configured by your team. It can tell you an accrual is due; it cannot build the accrual, because it does not know your proration rules or your vendors.
We build AI agents that run in your cloud, against your invoices and your GL, on rules encoded from your close calendar. You own the IP; proposed accruals reach a person with the working attached, and only approved entries post.
We need to get our books closed quicker. This would help shave a substantial amount of time.
Led by Peter Enestrom.
Asked about financial close automation.
The straight answers, before you book anything.
Month end close automation is using AI agents to do the repetitive work of closing the books: prorating invoices that cross the close boundary, proposing AP accruals, deduping related vendor bills, and drafting the journal entries. The judgment stays with your team — agents propose, staff review and approve, and only approved entries post to the GL.
AP accruals are the expenses you record at month end for bills you have not received or posted yet, so the month reports honestly. At a multi-branch company that means finding every invoice that crosses the boundary, splitting it across the right periods, and deduping the related vendor bills — by hand. One CFO told us it takes two or three days of their close period.
AP invoice automation is the pay side: capturing, matching, and approving the bills you receive, all month long. This page is the close side: getting the books closed at month end — accruals, proration, dedupe, entries. The two workflows share data and we build both, but they solve different pains. The pay-side workflow is covered on our AP invoice automation page.
Automated accruals are accrual entries an AI agent drafts instead of a person keying them. The agent prorates each boundary invoice against your close calendar, applies your thresholds, and proposes the entry with its working shown. A person approves it — human in the loop by design — and approved entries batch-upload into Sage Intacct or whatever GL you run.
Keep it — reconciliation software is a good checklist. What it cannot do is clear the items on the checklist: it tracks that an accrual is due, but a person still builds it. AI agents do the building, so every reconciling item that follows a rule gets worked automatically, and the software keeps doing what it is good at: tracking status.
The buyer who brought us this pain was losing two or three days of the close to AP accruals alone. Your number depends on how many invoices cross your boundary and how many entities you close — tracing one close cycle puts a figure on it before anything gets built.
Start with one workflow.
Tell us where your team loses hours. We will come back with a straight answer on whether AI can help, what it would take, and what it would pay.