The Fragmented-Systems Tax: Why Manual Excel Workflows Are Still the Default

Ask most GPs whether they've adopted AI and the answer is yes. Ask what their data infrastructure looks like underneath that AI and the answer, for most firms, is Excel.

Allvue's 2026 GP Outlook Survey, based on interviews with 102 private equity, private credit, and venture capital firms across the US, Canada, UK, and Benelux, found that firms with higher data maturity and AI adoption report stronger operational outcomes and better performance visibility. It also found that despite recognizing this, most GPs still rely on fragmented systems and manual, Excel-based workflows, and that system integration gaps remain the primary obstacle to automation and AI readiness. One respondent put it plainly: too much time goes to manual processes that should already be automated, and the real AI opportunity only begins once the underlying data is organized.

That is the fragmented-systems tax, and it is paid twice. Once in the obvious way: analyst hours spent reconciling spreadsheets that should already talk to each other. Once in a way most firms haven't priced yet: every AI tool layered on top of that fragmentation inherits the fragmentation. A due diligence copilot pointed at a data room is genuinely useful. A due diligence copilot with no connection to the fifty prior deals sitting in disconnected folders, systems, and inboxes is solving a narrower problem than the fund thinks it is.

This is the same failure mode MIT's NANDA research identified across enterprise AI broadly: tools stall not because the model is weak but because the organization underneath it cannot feed it anything durable. In private equity specifically, this shows up as a ceiling on exactly the workflows LPs care most about. Quarterly reporting cycles that could run in ten business days instead of six weeks stay at six weeks, because the underlying data still has to be assembled by hand before any AI layer can touch it.

The fix is not another point tool. It is treating institutional data, precedent, and judgment as infrastructure that compounds, the same way a firm treats its balance sheet or its LP relationships. Funds that make that shift first will not just move faster. They will be the only ones who can show an LP, credibly, that this year's diligence process is sharper than last year's because it remembers last year.

The tax gets paid whether or not a fund notices it. The only choice left is whether to keep paying it.