Five ERPs, one report: why the 13-week cash flow dies in Excel

Five acquisitions, five ERPs, one Excel tracker holding the fund’s weekly cash picture. The model does not die from hard math. It dies because somebody re-types it.

Monday morning at a lower-mid-market fund: the operating partner opens the portfolio tracker, the one they built themselves. Seventeen companies — or five, or nine — each on its own ERP, one of them a 14-year-old vertical system held together by manual workarounds. The week’s cash position lives in BAI bank files, portal logins, and a board pack each portco emails when it gets around to it. The tracker is the weekly liquidity view your 13-week model is trying to be. It is maintained by hand, and it dies a little every week.

Why the 13-week dies in Excel

The model is not the hard part. The hard part is that every input has to be re-keyed: bank balances from portals, cash summaries from PDFs, the portco CFO’s own spreadsheet with its own definition of "available." Every acquisition multiplies the systems. A week where one company is late turns the portfolio number into a guess, and the guess is what goes into the deck. Re-keying also freezes the format — change a threshold or add a covenant view and you are editing seventeen tabs by hand, so nobody does. The model stays frozen at whatever the first version assumed.

How do you get one report out of five ERPs?

You do not integrate the ERPs — you read the bank feeds. Every company’s transactions clear a bank, and bank data arrives in the same shapes — BAI files, direct connections — whatever ERP produced the entries. The build is a mapping, done once per company: which accounts are operating cash, which outflows are debt service, which thresholds count as late or unusual for that business. After that the weekly position assembles itself. Feeds land on schedule, the fund’s monitoring rules run against them, and anything the rules cannot reconcile — a missing feed, an unusual outflow — goes to a person in an exception queue with the reason attached. No portco installs anything, no portal gets adopted, and the next acquisition adds one mapping, not one more system to standardize. The report stops depending on who had time to type it.

The alternative everyone has tried is the portal: buy the platform, ask every portco to report into it. It dies at adoption — seventeen companies on seventeen ERPs are seventeen reasons the portal never gets updated, and the platform can only hold what gets typed in. The fund ends up back in the operating partner’s Excel, which is where it started, now with a license fee.

What the manual week actually costs

We have measured the assembly, not the analysis: recurring investor and portfolio decks hand-built at 10–12 hours of executive-assistant time each, so the firm’s picture of its own companies runs a cycle stale. And the tracker has a bus factor of one. When the person who built it leaves, the weekly cash position leaves with them.

Monitoring that asks seventeen portcos to update a portal dies at adoption. Reading beats asking.

This is implementation work, not a software category. Zaigo is an AI implementation firm, not a platform vendor: we encode the fund’s monitoring rules — account maps, thresholds, covenant definitions — against the bank feeds each portco already has, then run the encoded report beside the manual tracker for the first cycles, reconciled line by line before anyone relies on it. The full engagement shape is on the portfolio monitoring page (zaigo.ai/workflows/portfolio-monitoring); how we work with sponsors across a portfolio is on the PE-backed page (zaigo.ai/who/pe-backed).

If the weekly cash picture currently depends on one person’s spreadsheet, the next step is a 30-minute working call: zaigo.ai/book-a-call. We will tell you on the call whether the bank-feed route fits your portfolio — and what the first cycle costs.