Ask most finance teams how long it takes to produce the monthly board pack, and the honest answer is measured in days, not hours. Someone pulls extracts from the core system. Someone else reconciles them against the ledger in a spreadsheet nobody else fully understands. A third person rebuilds the same three charts they rebuild every month. By the time the deck lands in the CFO’s inbox, the numbers describe a version of the business that stopped existing four days ago.
For years that was simply the cost of doing finance. In 2026 it’s starting to look like a liability — and the firms that have noticed are quietly rebuilding the plumbing underneath their reporting.
The timing isn’t accidental. Deloitte’s inaugural Finance Trends 2026 report found that finance leaders now rank data privacy among their top three operational risks, close behind economic uncertainty and financial reporting itself. Nearly two-thirds of finance leaders plan to add technical skills like AI, automation, and data analysis to their teams over the next two years. The direction of travel is unmistakable: finance is being asked to move faster, govern tighter, and think more strategically — and you cannot do any of that on a four-day reporting lag.
Why the manual model quietly breaks
The spreadsheet-and-extract approach doesn’t fail loudly. It degrades.
Every manual handoff is a place where a number can drift. Every analyst maintains their own copy of “revenue,” each with a slightly different filter, so two reports that should agree don’t — and the finance meeting turns into a debate about whose figure is right instead of what the figure means. Reconciliation eats the first week of every month. And because the whole process is human-paced, the organisation can only ever look backwards. You report the quarter that ended; you don’t steer the one you’re in.
There’s a governance cost too, and it’s the one keeping heads of data awake. When data lives in email attachments and desktop files, nobody can answer the question a regulator or auditor actually asks: where did this number come from, who touched it, and can you prove it. In a sector facing overlapping mandates — operational resilience under DORA, risk-data aggregation under BCBS 239, financial reporting controls, privacy rules — “it’s in Priya’s spreadsheet” is not an audit trail. It’s an exposure.
The case: a specialty lender goes from four days to hours
Consider a mid-market specialty lender — the kind of firm with a few billion in its loan book, a lean finance team, and a board that wants sharper answers faster. Their month-end looked like everyone else’s: four business days to close and produce management reporting, most of it spent moving data between the loan-origination platform, the general ledger, and a stack of Excel models.
The rebuild didn’t start with a new dashboard. It started with a new foundation.
They consolidated their sources — loan servicing, the ledger, payments, the CRM, and their KYC provider — into a single governed data platform on Microsoft Azure, using Microsoft Fabric as the backbone and Power BI as the reporting surface. Instead of nightly extracts landing in someone’s inbox, data now flows continuously into one storage layer, gets shaped through defined stages, and surfaces in dashboards that refresh close to real time.
The result wasn’t incremental. Month-end reporting that took four days now takes a few hours, and the numbers on the board’s screen reflect the business as it is this morning, not as it was last week. More importantly, the finance team stopped spending its first week reconciling and started spending it analysing — which is the shift Deloitte is really describing when it talks about finance’s expanding remit.
That’s the headline outcome. The architecture that makes it repeatable is worth understanding, because it’s the same pattern showing up across the sector.
The architecture pattern, in plain terms
You don’t need to be an engineer to grasp why the new stack is faster. You need to understand three moves.
One copy of the data, not twenty. The old model copied data everywhere — into each report, each model, each analyst’s machine. The Fabric approach lands everything once in a shared store (Microsoft calls it OneLake) in an open format, and every report reads from that single copy. Governance, lineage, and security policies then apply uniformly from the raw data all the way through to the published report, closing the gap that appears when reporting operates under separate rules from the warehouse. One copy means one definition of “revenue,” one version of the truth, and no more meetings spent arguing about whose spreadsheet is correct.
Data is refined in visible stages. Raw data lands, gets cleaned, then gets shaped into business-ready tables — a progression teams describe as bronze, silver, gold. The value for a finance leader is that this pipeline is inspectable. When a number looks wrong, you can trace it back through each stage instead of guessing. That traceability is exactly what turns a reporting process into an audit-ready one.
Dashboards read the data live, without copying it again. The old trade-off in Power BI was brutal: either import the data for speed and accept that it’s stale, or query it live and accept that it’s slow. Direct Lake mode removes that compromise — Power BI reads directly from the shared store with no import step, giving near-instant dashboards over large volumes of data. This is the technical reason “four days to hours” is even possible. The reporting layer is no longer waiting on an overnight refresh; it’s reading the current state.
Stitch those three together and the four-day close doesn’t get optimised — it gets designed out.
Compliance stops being a fire drill
Here’s the part that turns a data project into a board-level one: in this architecture, governance isn’t a control you bolt on afterward. It’s part of the foundation.
Because everything flows through one platform, a governance layer — Microsoft Purview, in the Azure stack — sits across the whole estate. It tracks data lineage, classifies sensitive fields, applies sensitivity labels, and enforces permissions consistently from the landing zone through to the final report. When an examiner asks where a figure came from, the lineage is already recorded. When a privacy rule requires you to know exactly where customer PII sits, the classification already flags it. Identity and access run through Microsoft Entra, so who-can-see-what is enforced centrally rather than per spreadsheet.
The practical effect is that regulatory reporting shifts from a quarterly scramble to something closer to a standing capability. DORA’s expectation of rapid incident detection and BCBS 239’s demand for continuous — not audit-week — data integrity both assume you can see and prove the state of your data at any moment. That’s a reasonable ask when governance is engineered in. It’s nearly impossible when your controls live in a binder and your data lives in inboxes.
For a CFO or CIO, this reframes the whole investment. You’re not buying faster charts. You’re buying down regulatory risk and the personal accountability that increasingly comes with it.
KYC and the real-time risk picture
The same foundation quietly upgrades financial-crime controls. In the old model, KYC and AML data lived in a separate system that finance and risk touched only when they had to. Onboarding checks happened at the edge; the results rarely made it back into any unified view.
When the KYC/AML provider feeds into the same governed platform through a connector or API, customer risk data joins everything else the firm knows about that relationship — exposures, transaction patterns, servicing history. Risk teams get a current, connected view instead of a quarterly snapshot, and the sensitive fields that come with KYC data are classified and protected the moment they land, rather than tracked in a side process. Onboarding friction drops because the data is already where it needs to be, and the firm can watch for the patterns that matter continuously rather than in retrospect.
None of that requires ripping out the specialist KYC tooling. It requires giving it somewhere governed to land.
What actually changes for the finance leader
Strip away the architecture and the shift is simple to state. Finance moves from reporting the past to steering the present.
When the close takes hours instead of days, the CFO can ask “what if” questions and get answers inside the same meeting — which is precisely the advanced scenario planning Deloitte found leaders scrambling to build. In its survey, strengthening scenario planning and adopting more agile governance to support faster decisions were the two most common responses to macro uncertainty. A firm that can re-forecast in an afternoon navigates a rate shock or a credit wobble very differently from one waiting on next month’s pack.
It also lays the groundwork for everything coming next. A governed, unified, real-time data platform is the prerequisite for credible AI in finance — forecasting, anomaly detection, automated commentary. Bolting AI onto four fragmented spreadsheets produces confident nonsense. Running it on a single governed source produces something a board can trust. The firms rebuilding their infrastructure now aren’t just fixing month-end; they’re clearing the runway for the analytics they’ll be expected to have in two years.
The window is closing on “we’ll get to it”
The uncomfortable truth in the Deloitte data is that this is no longer a leading-edge move. More than half of finance leaders now say they play a lead role in shaping enterprise strategy — a seat you can’t hold while your team spends the first week of every month reconciling extracts. The firms that have rebuilt their data foundation are making faster decisions, passing audits with less drama, and freeing their best people from spreadsheet janitorial work. Their peers are still closing the books on Thursday for a month that ended on Friday.
That gap compounds. Every quarter a firm runs on manual reporting is a quarter its competitors spend building the muscle — and the trust in their own numbers — that real-time infrastructure makes possible.
Going from four days to hours isn’t really a story about speed. It’s a story about what finance gets to do with the time it wins back.
Rebuilding a financial firm’s data foundation is as much a sequencing and governance decision as a technology one — the order you migrate sources, and how you build compliance in from the first stage, largely determines whether the result is audit-ready or just faster. That first architectural decision is worth getting right before the first pipeline is built.