The month-end close is a ritual. Most finance teams treat it as a fixed feature of the month, with predictable stress, predictable hours, and predictable late nights. The team we worked with had been closing on day five for years. They knew the rhythm. They had also, every year, talked about doing it faster and never quite succeeded. The conversation that broke the pattern was not about technology. It was about what month-end actually is.
The realization that changed the design
Month-end close is, mathematically, a deadline-driven batch of reconciliations that have to be done anyway. The question is not whether the reconciliations happen but when. Doing all of them in the first five days of the new month creates a backlog that did not exist on day twenty-five of the previous month, when the team had spare capacity. Deloitte's 2024 Finance Transformation Survey found that more than 80 percent of finance professionals spend most of their close cycle on manual reconciliations and data preparation rather than analysis or insight, and organizations using accounting automation can improve speed to close by up to 50 percent by automating reconciliations throughout the month [1]. The unlock is not automation per se. It is the recognition that reconciliation work can be distributed across the month, which automation makes feasible at scale.
The reconciliations have to happen anyway. The question is not whether. The question is when.
What the team automated
Three workflows moved from month-end to continuous. Bank reconciliation, run daily against feed-driven statements, with exception routing for items the rules could not match. Sub-ledger to general-ledger reconciliations, run weekly, with variance reporting that escalated only when thresholds were breached. Intercompany matching, run daily for the largest entities and weekly for smaller ones. Each of these had previously consumed two to three days in the first week of close. After redistribution, each consumed roughly 20 to 30 minutes of attention per cycle, spread across the month, with the actual work done by rules and the human attention reserved for exceptions.
What the team did not automate
Important to be specific about what did not change. Manual journal entries for management adjustments, accruals requiring judgment, and consolidation overlays remained human-led. The automation did not touch these because the value of human judgment exceeded the value of automation efficiency at this layer. The right test is always: what is the cost of getting this wrong versus the cost of doing it manually. For high-volume reconciliations, automation is the obvious answer. For low-volume, high-judgment entries, manual is the obvious answer. The mistake most close-acceleration projects make is automating both layers and creating governance problems in the second one.
The first three closes after the change
Close one took six days, slightly longer than the historical baseline. The team was learning the new rhythm and the exception routing was tuned too sensitively. Close two took four days, faster than baseline but not yet at target. Close three took three days and has held there since. The pattern is consistent with what we see across mid-market finance transformations: the first close after a workflow change is slower, not faster, because the team is debugging the new process while also doing the close. Setting expectations for this beforehand is more important than the technology choice.
The first close after the change is always slower, not faster. Setting that expectation before the change is the most valuable thing the project lead does.
The headcount question
One FTE left the team between the start of the project and close eight. The departure was not driven by the automation; it was a planned career move. The team did not backfill, and the close held at day three with the smaller team. This is the honest version of 'headcount savings from automation' in finance. The technology did not displace anyone. It made the team capable of running the close with one fewer person when natural attrition created the opportunity. Most finance automation savings emerge this way. The framing of headcount displacement misses the actual mechanism, which is capacity for next year's growth without next year's hiring.
What the CFO got back
The two days saved on the close show up in two places. First, the FP&A team starts variance analysis two days earlier, which means the monthly management report reaches the executive team in week one rather than week two. Second, the controller and senior accountants spend the recovered time on substantive work that had previously been deferred indefinitely: process documentation, control testing, ad hoc analysis for the business. The CFO's subjective experience of finance changed more than the headline metrics suggest. The team was no longer in close mode for half the month.
The governance trade-off
Continuous reconciliation creates an audit trail that is denser and more granular than monthly batch. This is mostly a benefit but it introduces a governance consideration: more reconciliation events means more change records, which means external auditors have more to sample. The audit firm working with this team adapted by moving to data-analytics-driven testing on the continuous data, which Deloitte's 2024 audit transformation work specifically supports through agentic AI in the Omnia platform [2]. The audit relationship improved, not because of the technology, but because the conversation became about the data itself rather than about reconciling differences between the client's records and the audit team's reconstruction.
For Belgian mid-cap finance teams specifically
Two factors make this approach particularly viable in the Belgian mid-cap context. First, the level of system consolidation in mid-market Belgian firms is generally higher than in larger international groups, which means the data feeds required for continuous reconciliation are more achievable. Second, the talent market for senior finance roles is constrained, which makes capacity-multiplying investment more valuable than capacity-replacing investment. The teams that have built continuous-close architectures are better positioned to absorb growth without growing the team, which matters in a market where finance talent is the binding constraint.
