AI Agents for the Sage
Intacct Close
A small subsidiary with one part time bookkeeper can hold a group close for a week, and usually gets blamed for more than it caused.
- The reason teams choose Intacct is multi entity, and that turns a short close into a sequence repeated per entity plus a consolidation blocked until the last one finishes.
- In a worked group close of nine entities running three days late, two and a half days sat at group level or between the two largest entities. The small entity everyone blames accounted for one.
- Measure where the close actually waits before pointing a process improvement effort at it. The answer usually contradicts what everyone believes.
- Reconcile intercompany at transaction level. Two errors cancelling across a pair pass a totals review and leave both entities misstated.
- Seven per entity checks are worth running every period, all are easy, and all get skipped when the close is late, which is exactly when they matter.
AI agents for the Sage Intacct month end close run the per entity checks in parallel rather than in sequence, reconcile intercompany at transaction level from both sides at once, and report which entity is actually holding the consolidation. ChatFin prepares every correction as a proposal for a named person to release.
The multi entity structure is why teams choose Intacct and it is also what makes the close long. Reconciliations multiply by entity, intercompany runs every pair every month, currency needs revaluation and translation, and the consolidation cannot start until the slowest entity finishes. That last constraint is what makes one entity everybody's problem, and it is rarely the entity everyone assumes.
The close is multi entity, and that is the whole difficulty
A single entity Intacct close is quick. The reason teams choose Intacct is multi entity, and that turns a short close into a sequence repeated per entity plus a consolidation that cannot start until the last one finishes.
| Step | Single entity | Multi entity |
|---|---|---|
| Reconciliations | A handful | Times the number of entities |
| Intercompany | None | Every pair, every month |
| Currency | Usually none | Revaluation and translation, per entity |
| Allocations | Simple | Across entities, with a basis to defend |
| Consolidation | Not applicable | Blocked until every entity closes |
The consolidation being blocked is what makes one slow entity everybody's problem. A small subsidiary with one part time bookkeeper can hold the group close for a week, and usually does.
ChatFin runs the per entity checks in parallel rather than in sequence and reports which entity is holding the consolidation and why, which is usually the most useful single output of the whole close.
Intercompany, reconciled at transaction level
Mechanical, recurring and always late. The reason is that each side is drawn separately and the totals are compared.
Draw both sides at once
One read, both entities, matched at transaction level. Reconciling at totals lets two errors in opposite directions net to something small that passes review.
Name the document and the period on both sides
Most differences are one entry booked in a different period. Saying so ends the exchange in a message rather than a thread.
Report recharges with no supporting document
These are the ones that become an argument at year end, and they are findable now.
Check the allocation basis is the documented one
Cross entity allocations drift. Recomputing under the written basis and reporting the difference is mechanical.
A worked group close, 9 entities
Nine entities, consolidation due on working day 6, delivered on day 9. The delay is always attributed to the two smallest entities. Reading the actual close timings says something else.
| Cause of delay | Days | Where it actually sat |
|---|---|---|
| Intercompany differences unresolved | 1.5 | Between two large entities, not the small ones |
| One entity's bank reconciliation | 1.0 | The small entity everyone blames |
| Revaluation rerun after a late rate | 0.5 | Group, avoidable with a rate cut off |
| Allocation basis queried | 0.5 | Group, recurring every quarter |
| Waiting for a manual adjustment schedule | 0.5 | One person, undocumented |
Two and a half of the three days sit at group level or between the two largest entities. The small entity everyone chases accounts for one day. That is worth knowing before the next process improvement effort is pointed at the wrong place.
The per entity checks worth running every period
Run them per entity, in parallel, and report only exceptions.
| Check | What it catches |
|---|---|
| Control accounts tie to subledgers | A journal posted straight to a control account |
| Bank reconciled item by item | Unrecorded charges and missing receipts |
| Dimension completeness on the period's postings | Reports that will be quietly wrong |
| Recurring entries that did not change | A copy forward nobody reviewed |
| Manual journals with no reference or support | The audit finding waiting to happen |
| Postings dated outside the period | Cut off errors, both directions |
| Intercompany balanced at transaction level | The offsetting pair that nets to nothing |
None is difficult. All get skipped when the close is late, and the close is late precisely when they matter most.
ChatFin runs all seven per entity on one read, so they cannot disagree with each other the way separately built reports usually do, and reports which entity each exception belongs to.
Where this underperforms on Sage Intacct
Three.
Deeper in the Sage Intacct close
Three use cases sit under this hub, in build and shown rather than linked.
Worked on this page
Close task monitoring · Close exception identification · Flux analysis
Covered as sections here
Intercompany · Multi entity consolidation · Control account tie out · Bank reconciliation · Cut off · Allocations
Questions Sage Intacct close teams ask first
Which entity is really holding up a group close?
Usually not the one being blamed. In a worked close of nine entities running three days late, 1.5 days went to unresolved intercompany differences between two large entities, 0.5 to a revaluation rerun after a late rate, 0.5 to a queried allocation basis and 0.5 to waiting on one person's undocumented adjustment schedule. The small entity everyone chases accounted for one day. ChatFin measures where the close waits, per entity.
Why does intercompany take so long every month?
Because each side is drawn separately and the totals are compared, which turns into an email thread between two teams holding two different extracts. Most differences are one entry booked in a different period. ChatFin draws both sides at once and reconciles at transaction level, naming the document and the period on each side, which also catches the offsetting pair of errors that nets to almost nothing and passes a totals review.
Which checks are worth running every period?
Control accounts tied to subledgers, bank reconciled item by item, dimension completeness on the period's postings, recurring entries that did not change, manual journals with no reference, postings dated outside the period, and intercompany balanced at transaction level. None is difficult and all get skipped when the close is late. ChatFin runs all seven per entity on one read so they cannot disagree with each other.
Find out which entity is actually holding the group.
ChatFin runs the per entity checks in parallel, reconciles intercompany at transaction level from both sides, and reports where the close is genuinely waiting rather than where it is assumed to be.
Every correction arrives as a proposal with its evidence attached, for a named person to release.
Bring one group close. The timing analysis takes an hour.
Book a demoAll images are served from the Pexels content network. No third party product screenshots, vendor logos or licensed media are used on this page.
Bring one group close.
We will measure where it actually waited through ChatFin, per entity, and reconcile one intercompany pair at transaction level.