Virtual data rooms — Intralinks, Datasite, Ansarada, Box, Droppoint — have transformed how deal teams share documents during M&A due diligence. They solved a real problem: securely distributing thousands of documents to multiple buyer parties with granular permission controls and audit trails. That's a genuine operational improvement over the physical data rooms of the 1990s.
But here's what VDR vendors don't advertise: their platforms are document repositories, not analysis tools. They tell you who accessed a file and when. They do not tell you whether what's in the file matches the financial ledger. That gap — between 'documents are accessible' and 'documents are verified' — is where deals go wrong.
What VDRs actually do
VDRs are excellent at three things: document storage, access control, and Q&A workflow management. They give sellers a controlled environment to share sensitive documents and give buyers a structured way to submit diligence requests and track responses. The audit trail feature — who viewed what and for how long — provides a record that's valuable in post-close disputes about disclosure.
Most enterprise VDRs also offer bulk download, full-text search, and some document organization features. A few have added AI-powered document review layers. But even the most advanced VDR platforms stop at the document level. They surface text within a document. They don't cross-reference that text against a financial ledger.
The VDR knows that a vendor invoice was uploaded on March 12th and viewed by the buyer's counsel for 4 minutes. It doesn't know whether the invoice amount matches the corresponding accounts payable entry — or whether the invoice's reference number appears anywhere else in the room.
The analysis gap that VDRs leave open
After a buyer's team has access to a VDR, the actual analysis work falls entirely on the deal team. Someone — typically a first or second-year analyst — downloads the financial ledger, downloads the supporting documents, and manually reconciles them. The VDR's audit trail records the downloads. It doesn't help with what happens after.
This is the analysis gap: the space between document access and document verification. In that gap sits the real diligence work — and it's still almost entirely manual at most firms. The VDR is infrastructure. Analysis is not.
What happens in the analysis gap
- —Analysts download batches of documents and open them one by one
- —Reference numbers and amounts are manually compared to ledger rows
- —Discrepancies are logged in separate Excel tracking sheets
- —Coverage is inconsistent — some documents are reviewed thoroughly, others skimmed
- —The output is a color-coded spreadsheet with no machine-readable structure
Why VDR vendors haven't closed this gap
VDR vendors are in the access-control business. Their value proposition is secure document distribution to multiple parties with sophisticated permission hierarchies. Adding a reconciliation layer would require ingesting and understanding the buyer's financial ledger — a document the seller doesn't have and the VDR vendor has no business relationship with. The gap isn't an oversight; it's structural.
Some VDRs have added AI features — document summaries, smart search, risk tagging. These are useful for initial document triage, but they operate document-by-document. They don't answer the question 'does this document support this ledger entry?' They answer 'what is this document about?' Those are different questions.
What closing the gap requires
Closing the analysis gap requires a tool that operates on both sides of the equation: the financial ledger on one side and the data room documents on the other. The core technical problem is extraction plus matching — extracting structured data from heterogeneous document formats (PDF invoices, Excel schedules, CSV exports, scanned bank statements) and matching that data against ledger rows with enough intelligence to handle formatting variation, typographic errors, and semantic equivalence.
STET was built specifically for this problem. It connects to Box and Dropbox VDRs, ingests the financial ledger, and runs a multi-pass matching pipeline that starts with exact matching and escalates to semantic ML only for the cases that require it. The output is a discrepancy report with evidence links to specific pages in specific documents — not a broad AI summary, but a specific finding with a specific location.
The practical implication for deal teams
VDR selection matters for deal process management. But it doesn't determine the quality of your diligence analysis. A well-organized VDR with a rigorous manual reconciliation process produces better outcomes than a poorly-organized VDR with automation — but the ceiling on manual reconciliation is real.
The analysts who run reconciliation on large deals know this. They've seen the 3 AM spreadsheets. They know that coverage at scale is genuinely difficult without a different approach. The VDR is necessary infrastructure. It's not sufficient analysis.