Back to Blog Listing

Real-time treasury: the mid-market data integration gap

Real-time treasury: the mid-market data integration gap
Agata Wojtas Jul 11, 2026 4 min read

Written by: Agata Wojtas, Chief Commercial Officer, Digital Colliers

Your treasury dashboard says one thing. Your bank portal says another. Your PSP settlement report will say a third thing tomorrow morning. If you run treasury at a mid-market financial services firm, you already know the dashboard is a lagging indicator dressed up as a live one. The question is how much that gap is costing you, and how far ahead the operators who've closed it have pulled.

The lag is structural, not lazy

The lag isn't because your team is slow. It's because the data pipe was never built for real time.

Most mid-market finance teams still run month-end close in spreadsheets, pulling numbers across systems by hand. That same pattern repeats itself daily inside treasury. Someone exports a CSV from the core banking portal. Someone else pulls the card processor settlement file. FX positions get reconciled against yesterday's rates. By the time the dashboard refreshes, the underlying reality has moved.

The usual culprits, in order of how much pain they cause:

  • Bank feeds that arrive on a T+1 batch cycle rather than via API
  • Card processors and PSPs where settlement and reporting run on separate clocks
  • FX providers whose rate feeds and trade confirmations live in different systems
  • Intercompany positions that only reconcile at month-end
  • Manual GL postings that don't hit the warehouse until someone runs the job

Each one is defensible on its own. Stack them and your "cash position" is a polite fiction.

What this costs on a Tuesday afternoon

The close-cycle numbers give you the clearest picture of the gap. Month-end close at mid-market finance teams typically runs 8 to 10 days. The teams that have actually invested in integrated data close in under 5. That's not a productivity difference. That's a decision-latency difference.

Translate that to treasury and it looks like this. You're deciding whether to draw on a revolver, sweep cash to a higher-yield account, or hold back on an FX hedge. You're making that call against a picture that's anywhere from 18 hours to 4 days behind. In a rising-rate or volatile-FX environment, a two-day-old cash position is a guess. If you guess wrong, you either park cash you should've deployed or draw credit you didn't need.

The risk isn't just yield drag. Under DORA, which has been in force across the EU since January 2025, operational resilience includes the integrity and timeliness of your financial data flows. Regulators are increasingly asking not just whether your reports are accurate, but whether you can produce them on demand. A ten-day close cycle is not a good answer to that question.

What live integration actually looks like

The operators who've fixed this haven't bought a product. They've rebuilt the pipe.

The pattern I keep seeing has four pieces:

  1. Direct API connections to every bank, PSP, and FX venue, with a fallback to SFTP where the bank hasn't caught up. No manual CSV pulls.
  2. A single normalised transaction store, with each event timestamped when it happened at the source, not when it landed in the warehouse.
  3. Reconciliation logic that runs continuously, not nightly. Breaks are flagged in minutes, not at month-end.
  4. A dashboard layer that reads from that store, with visible freshness indicators so the treasurer knows the age of every number on screen.

None of that is exotic engineering. What it needs is senior people who can navigate the messy edges: bank formats that don't match spec, PSP webhooks that occasionally drop, FX confirmations that arrive out of order. That's where most of these projects fail. Around 95% of enterprise AI and data projects don't reach production or ROI, and the ones that die inside treasury usually die on integration, not on the model.

The left-behind risk

Here's the uncomfortable part. The gap between mid-market and best-in-class close cycles isn't closing. It's widening.

Firms that got their treasury data flow right two years ago now compound the advantage. They deploy cash faster, hedge earlier, satisfy DORA reporting requests inside a day, and free their finance leadership from reconciliation firefighting. The firms still running yesterday's dashboard are quietly falling behind on yield, on capital efficiency, and on the regulator's patience.

You don't need a full replatform to start. You need to know which of your feeds is oldest right now, and what a real-time version of that one feed would change on Tuesday afternoon. Start there.

Related Posts