Written by: Luke Sobieraj, Founder & COO, Digital Colliers
The board asks the same six questions every month. Revenue by segment. Margin trend. Cash position. Pipeline coverage. Headcount versus plan. Regulatory exposure. And somewhere in the finance org, an analyst spends most of the week stitching the answers together in Excel.
The questions don't change. The answers do. That gap, the one between a stable question set and an unstable assembly process, is where most mid-market finance teams are quietly bleeding capacity. It's not a reporting problem. It's a data integration problem wearing a reporting costume.
The pattern is the same everywhere
Walk into ten mid-market finance teams and you'll see the same choreography. Someone exports from the ERP. Someone else pulls the CRM extract. A third person owns the HRIS numbers. A fourth reconciles the treasury feed. Then a senior analyst joins them all in a workbook that only they fully understand.
This is not a caricature. FSN Research has found that most mid-market finance teams still run month-end close in spreadsheets, pulling numbers across systems by hand, and that close typically takes eight to ten days. The teams considered strong at this close in under five. The delta is not talent. It's whether the joins between systems live in a person's head or in code.
The board pack sits downstream of all of this. So every question the board asks quietly re-triggers the same manual join. Revenue by segment means reconciling the CRM's account hierarchy against the ERP's revenue lines. Regulatory exposure means walking three systems and a shared drive. Nobody wrote it down because the person doing it already knew.
The cost of leaving it manual
Here's what the cost of inaction actually looks like, once you count it honestly.
- Four days a month of senior analyst time, repeated forever, on work that produces no new insight.
- A close cycle that runs eight to ten days when the strong benchmark is under five, which means the board is looking at data that's already two weeks stale.
- Key-person risk on the one analyst who knows how the workbook actually ties out.
- No audit trail worth the name, which matters more every quarter as DORA obligations bite. DORA has been in force across the EU since 17 January 2025, and it expects you to know where your numbers come from.
- Zero reusability. When the board adds a seventh question, you start from scratch.
And this is before anyone asks the AI question. Roughly 95% of enterprise AI projects fail to reach production or ROI, and a large share of those failures trace back to the same root cause: there's no clean, governed data layer for the model to sit on top of. If your humans can't answer the six board questions from a single source, your copilot can't either.
The semantic layer is the actual fix
The move that separates the teams shipping this well from the teams still exporting to Excel is building a semantic layer between the source systems and every downstream consumer, whether that's a board pack, a dashboard, or eventually an LLM.
In practice that means:
- Pick the twenty to thirty metrics the business actually runs on. Revenue, margin, ARR, churn, DSO, headcount, the six board questions. Not two hundred.
- Define each one once, in code, with the join logic and the filters explicit. Segment definitions live here. Currency conversion lives here. The fiscal calendar lives here.
- Point everything at it. The board pack, the FP&A models, the ops dashboards, the eventual chat interface. One definition of revenue, everywhere.
- Version it. Treat metric definitions the way engineers treat code, with pull requests and history, because the audit conversation is going to want that trail.
This is not glamorous work. It's plumbing. But it's the plumbing that decides whether the next three years of finance automation actually lands or joins the 95% failure pile.
Why the CFO's team wins at thirty minutes
When the board pack refresh drops from four days to thirty minutes, the interesting thing is not the time saved. It's what the analyst does with the other three and a half days.
They start answering the seventh question. The one the board didn't ask because they knew it would take another week. They start running the scenario the CFO wanted last quarter. They start noticing the margin drift in segment three because they finally have time to look. The finance team stops being a reporting function and starts being a thinking function.
That shift is the actual prize. The board pack is just the forcing function that gets you there.