The close has a structural problem. Here's what it actually is and why the solution isn't what most finance leaders think.
|
6–10 Business days lost to close every month |
67% Of finance leaders cite close as their #1 stress source |
30+ hours/month spent on manual reconciliations
|
Let's be honest about something. You've probably tried to fix the close before.
Maybe you hired more analysts. Maybe you invested in an RPA tool that automated a handful of repetitive tasks. Maybe you tightened up your close calendar, sent the chasers earlier, and ran a post-mortem after every painful period end.
And yet. Here you are. Still losing a week — or more — every single month. Still watching your best people disappear into reconciliation hell while the business waits for numbers that should have been ready two days ago.
Here's the uncomfortable truth: it's not a people problem. And it's not an effort problem. It's a structural problem. And structural problems don't respond to more headcount or harder work.
Month-end close isn't slow because your team is slow. It's slow because of what the process actually requires: reconciling data across multiple systems that were never designed to talk to each other, applying human judgment to hundreds of individual transactions, chasing approvals across teams with competing priorities, and writing commentary from scratch on numbers that have only just become available.
Every one of those steps is a handoff. And every handoff is a place where delays compound, errors multiply, and your best finance people end up doing work that is, frankly, beneath them.
Think about what that actually means across a year. If your close takes 8 days, your finance team spends roughly 96 days — nearly a third of the working year — in close mode. Not in business partnering mode. Not in strategic planning mode. In close mode.
The opportunity cost of that is enormous. And it compounds every quarter, every year, every time a good analyst decides they'd rather work somewhere their skills are actually used.
A decade ago, Robotic Process Automation was sold to finance teams as the answer. Automate the repetitive stuff. Free up your people for higher-value work. It was a reasonable promise.
And RPA delivered — on a narrow slice of the problem. Rule-based tasks, fixed data formats, predictable exceptions: RPA handles those well. The problem is that the code isn't mostly rule-based, fixed-format, or predictable.
Distinguish a genuine anomaly from a normal pattern shift caused by a product launch or seasonal swing
Draft a correcting journal entry, present it for human approval, and post it to the GL — in one unbroken flow
Generate a plain-English narrative explaining why P&L moved the way it did this period
Handle intercompany disputes where both sides have slightly different numbers, and someone needs to reason through which is right
Learn from previous close cycles and improve its own judgments over time
RPA can move data between systems. It can't think about data. And the close fundamentally requires thinking.
This is why finance teams that invested heavily in RPA often find themselves in the same place as before: the repetitive tasks are faster, but the hard parts-the judgment calls, the exceptions, the explanations — still land on your most experienced people at the worst possible moment.
It's easy to treat the slow close as a known cost of doing business. It's less easy to quantify what it's actually costing you.
Analyst overtime during close windows
Error correction — restatements, audit findings, and the rework they generate
Delayed management reporting, which means delayed decisions
Analyst attrition: talented finance professionals don't join a company to reconcile spreadsheets for a third of their working year
Opportunity cost: every hour spent in the close is an hour not spent on forecasting, scenario modelling, or business partnering
Organisational credibility: a finance function that takes 10 days to close is perceived — fairly or not — as a constraint on the business, not a strategic asset
Manual journal entry errors remain a leading cause of restatements and audit findings. The close isn't just slow — it's risky.
The close doesn't need more people doing the same things faster. It needs a fundamentally different approach to how the work gets done.
Specifically, it needs automation that can reason — that can look at a set of unallocated transactions and not just flag them, but identify the right cost centres, draft the correcting entries, surface them for human approval, and post them once approved. It needs automation that can handle exceptions, not just avoid them.
That capability now exists. It's called agentic AI, and it represents something genuinely different from the copilots, chatbots, and RPA tools that finance teams have been using up to now.
In the next post in this series, we'll get into exactly what agentic AI is, how it works, and why the month-end close is the perfect first place to apply it.
But if you'd rather see it working right now than read about it — that's understandable too.
Watch a full month-end close run by AI agents, in real time, with live approvals.
Email us at media@octanesolutions.com.au