Month-end is not an accounting requirement. It is a batching artifact: the shape finance work takes when reconciliation, matching and elimination are expensive human tasks that have to be saved up and done in 1 push.
Close management software, as a category, exists to make that push shorter. Checklists, task owners, status dashboards, sign-off workflows. Every product in it accepts the batch as permanent and offers to project-manage it better.
The batch is not permanent. At Oper Credits, consolidation across 3 entities used to take, in the finance manager's words, "half a week, a little more." It does not take a shorter time now. It does not happen. Consolidation posts as transactions happen, and the month-end assembly step it used to belong to no longer exists.
of manual consolidation at Oper Credits. Not faster. Gone
days to close at Tillo across 4 entities and 23 currencies, mid-transition
days to close 4 countries at All Gravy, with nobody working nights to get there
Why the batch formed
Almost none of a long close is accounting judgment. It is retrieval and repair.
Pulling exports from each entity. Matching payments to bills that arrived 3 weeks earlier. Chasing receipts for card spend nobody documented at the time. Drafting eliminations for intercompany transactions that posted in the second week. Reconciling bank statements against a ledger untouched since the last close.
Every one of those tasks was small when it appeared and expensive by the time anyone reached it. The cost was never the work. It was the delay between the event and the work. A payment matched the day it lands takes seconds. The same payment matched on day 26 requires somebody to reconstruct what a round number from a vendor with 4 open invoices was settling.
So the batch is not a process choice. It is what accumulates when the only thing that can do the work is a person, and the person cannot do it continuously.
What dissolves it
Agents do not accumulate. A payment arrives and matches against its bill within minutes. A receipt goes missing and the chase starts that day in Slack, not at month-end from a spreadsheet of exceptions. An intercompany transaction posts and its elimination drafts immediately. Bank reconciliation runs against live feeds rather than once, heroically, in the first week.
Each task is trivial in isolation. That is the whole mechanism. The close was only ever a mountain because 30 days of trivial tasks were left to pile into one.
At All Gravy, 500 to 600 Pleo card transactions a month used to be uploaded by hand, and a single missing receipt held up an entry while somebody tracked exceptions across 2 systems. That number is now zero: the transactions land in the ledger without anyone touching them. Nothing about that is a faster month-end. It is a month-end task that stopped existing.
"Our close used to be excruciating. Lots of systems, lots of bookkeepers, lots of places where the whole thing could stall."
Sebastian Sandorff Jacobsen, Head of Finance and Ops, All Gravy
The 5-day close is the transition, not the destination
Tillo closes 4 entities and 23 currencies in 5 days, down from 12. All Gravy closes 4 countries in 5. Ocean.io cut close time 60%. Alva Labs is targeting a same-day close. Omnio, currently at 15 days, is moving onto Light with the explicit aim of reaching 1.
Read those as a category shortening and the point is lost. They are snapshots of the same journey at different distances along it, and the direction is not "shorter." It is "absent." Each of those numbers falls because another task moved out of the batch and into the moment it belongs to, and the number keeps falling as long as there are tasks left to move.
Old model
Administer the batch
Close software assigns tasks, tracks status and chases owners through a week of retrieval and repair that was always going to happen.
Light's model
Dissolve the batch
Matching, chasing, eliminations and reconciliation run as transactions post. Nothing accumulates, so there is nothing to coordinate.
What is left
Judgment, on current books
Estimates and accruals that need a human decision, made against a ledger already reconciled rather than one being rebuilt.
What survives is real but small. Judgment calls, estimates, the accruals that genuinely need a person. Those still happen on a cadence, and they happen against books that are already current, which changes what the exercise is. The close stops being reconstruction and becomes confirmation.
The precondition
Continuous close has an architectural requirement, and it disqualifies most stacks that claim it. The agent has to see the whole transaction. If accounts payable lives in 1 tool, expenses in another and the ledger in a third, nothing can match, chase and post across the gaps, and the batching returns at every system boundary. A close that is continuous inside 3 systems and batched between them is a batched close.
This is why the shortest closes belong to the teams that consolidated their stack, not the teams that bought the best close tool. All Gravy replaced 7 systems with 1, across 4 ledgers, a card provider and 2 payment platforms. Oper Credits went from 8 systems to 3. Tillo replaced 8 ledgers with 1.
What the week is for
The number worth watching is not the close duration. It is what a finance team does with a week it did not have.
At Tillo, finance "stopped being a boring back-office function and became a key player in driving growth and direction," says Harriet Stewart, Head of Business Systems. At Alva Labs a board report that took days assembles in seconds, because an agent with read access to the ledger and the CRM builds it on request, which means the analysis happens while the numbers still matter. All Gravy doubled with a finance team of 3, and its bookkeeping partner spends 30 to 40% less time on the books than before.
The close was a workaround for a constraint that has lifted. Teams still buying software to administer it are buying a better way to run the workaround. The rest are watching the batch get smaller every month, and there is no version of this where it stops shrinking.
Read the close step by step, see how Light closes continuously, or book a demo.