You’re renting a filing cabinet by the page. It can’t read anything inside it.

Ask a fund COO what their data room costs and you'll usually get a pause, then a range.

The pause is the interesting part. Per-page pricing, per-GB overages, per-project licences, per-user seats that renew whether or not anyone logged in this quarter. Very few people can state the number cleanly, which is itself a fact about the category.

But the cost isn't the real problem. This is:

When the raise closes, the data room goes dark. The LPA, the DDQ responses, the audited statements, the subscription documents — everything the next three years of LP questions will refer back to — sits in a system your operations team stopped paying attention to the day the fund closed.

And the whole time, the thing holding your most valuable documents could not read a single one of them.

A data room isn't a product. It's a permission model.

Strip away the branding and every VDR does three things: hold files, control who sees which ones, and record who looked.

We didn't build a separate product for that, because we already had the hard part — a folder tree with permissions that flow downward, sitting under the same roof as your funds, investors, commitments and KYC records.

So the data room is a folder. A special one, with rules:

• Investors never see the root. They see only the sub-folders you have explicitly shared with them. An LP with access to "Q3 Update" cannot see that "Lead Investor Pack" exists — not "can't open it", can't see it.
• Access cascades. Grant a firm access to a sub-folder, tick "apply to files and sub folders", and everything beneath it inherits. One decision, one place, no per-file permission grid to maintain and get wrong.
• Sharing outward doesn't require an account. A prospect who has never logged in gets a secure view link to a single document — and that link still respects that document's download and print settings.

Data Room/
├── Q3 Investor Update/     → shared with all LPs
├── Lead Investor Pack/     → lead investors only
├── DDQ Responses 2026/     → shared per requesting LP
└── New Deal Preview/       → not yet shared with anyone

Control that survives the download

Two features here do more work than their size suggests.

Download and print are per-document decisions. Turn them off and the viewer gets a protected, watermarked copy rather than the original file. The document is readable and not portable — which is the actual requirement when you're sharing a term sheet with eleven prospects.

Every view is recorded. Each document carries a list of who opened it and when. That's an audit answer, and it's also an intelligence signal: the LP who has opened the LPA four times this week is at a different stage of the decision than the one who never opened it at all.

Signatures live in the same place as the documents

Subscription documents are the clearest case for why a bolted-on e-sign integration is worse than a built-in one: the document being signed is the fund's document, on the investor's record, in the folder the auditor will look in.

So e-signing runs inside the same document store, and a daily agent watches the whole pipeline.

The E-Sign Monitor sweeps every document in the entity that was sent for signature and reports two categories: requests sitting unsigned past your configured window (three days by default), and documents that expired, were cancelled, voided, or failed.

Note carefully what it does not do. It never touches a document. It doesn't resend, re-trigger, or nudge the signer. Its report says so explicitly. It tells your team what's stuck; a human decides what to do about it.

That restraint is deliberate. An agent that automatically re-sends signature requests will, on the day it goes wrong, email a hundred LPs a second copy of a document they already signed. The cost of that mistake is far higher than the cost of a person clicking resend.

⚙️ Under the hood: the documents are readable

Here's the capability a rented data room structurally cannot offer, no matter what it charges.

Our documents are read by AI into Document Q&A — structured question-and-answer facts extracted from the content. A KYC form becomes queryable fields. A financial statement becomes retrievable figures.

The important part isn't the extraction. It's what consumes it: those facts power the platform's support agents and AI rules. The document isn't just stored and searched — it participates. A compliance rule can check a value that only exists inside a PDF. An agent preparing a capital call can verify something the ops team would otherwise have to open a document to confirm.

That's only possible because the data room isn't a separate silo. It's the same document store the rest of the fund runs on, which means anything that can read your documents already knows what a capital commitment, a portfolio company, and an LP are.

A standalone VDR could add AI search tomorrow — and it would still be AI over a folder of PDFs, with no idea which investor, which fund, or which commitment any of them belong to. Context isn't a feature you add. It's an architecture you either have or don't.

	Rented VDR	Data room inside the fund platform
Pricing	Per page, per GB, per project, per seat	Part of the platform
Lifespan	The project. Goes dark at close	Permanent — it's your document store
Permissions	Its own model, maintained separately	The platform's own inheritance, one place
Knows what a capital commitment is	No	Yes
Documents readable by your agents	No	Yes — extracted facts feed agents and rules
E-signature	An integration	The same document, same record
Audit of who viewed what	Yes	Yes — on the record the auditor already opens

🔭 What we're building next: agents that work the data room

Everything below is roadmap, not shipped. It's where this goes, and I'd rather describe it honestly than imply it exists. If one of these would change how your raise runs, tell me — it moves up the list.

1. Data Room Curator. Checks each shared folder against the pack it's supposed to be — LP DDQ pack, prospect teaser pack, annual pack — and reports what's missing, what's older than your staleness threshold, and what's duplicated across folders. The question it answers: is the thing we shared with this LP actually complete?

2. DDQ Responder. An LP sends a 200-question due diligence questionnaire. The agent drafts answers grounded strictly in the documents already in your data room, citing the source document and page for every answer, and flagging every question it cannot support rather than composing something plausible. Same rule as our side letter agent: no citation, no answer.

3. Version Drift Monitor. Finds superseded documents still sitting in shared folders — last year's LPA, the prior quarter's report, a term sheet that changed after it was shared — and the same document existing in two folders at two versions. This is the error class that causes real damage, and nobody catches it by browsing.

4. Exposure Check. Runs before a folder is shared, not after. Scans for content that shouldn't leave: another LP's name or commitment amount in a sample capital account, bank details, personal data in an unredacted attachment. Flags it to a human — never auto-redacts.

5. Engagement Signals. Reads the view trail and turns it into a fundraise read: which prospects are actually working through the pack, which document everyone abandons halfway, which LP went quiet three days after opening the LPA. Feeds straight into the relationship record so the follow-up is timed off behaviour rather than a calendar.

📊 The impact

Before: a per-project VDR priced by the page, holding the documents your LPs will ask about for the next decade, in a system that expires — with a separate permission model your team maintains twice and an e-sign integration that lives somewhere else again.

After: one document store where access cascades from a single checkbox, every view is recorded, download and print are per-document decisions, signatures happen on the same record, and a daily agent reports everything stuck in the signing pipeline.

The question worth asking internally: if a prospect asked for your full DDQ pack this afternoon, how long until you could confirm — not assume — that what you shared was complete and current?

If the answer involves someone opening folders one by one, that's the gap. It's also exactly what roadmap item #1 is for.

What to take from this

See it on your own structure

If you're paying per page for a room that shuts down at close, the exercise worth doing is simple: list the documents your LPs have asked about in the last twelve months, and check how many are in a system your ops team still opens.

If that list is uncomfortable, DM me and I'll show you what the same folders look like when the fund's own agents can read them.