Frequently asked questions

Vurdera turns a deal folder into an investment committee workspace.

It processes the documents in that folder — CIM, model, management presentation, diligence reports, DDQs, call notes — and produces a structured investment committee memo in your firm’s format. Every factual claim drawn from the deal materials is cited to the document and page it came from.

Committee members get a read-only view where they can ask questions about the deal documents and see each answer traced back to its source, or see plainly that the documents do not contain one.

Vurdera handles the document-heavy preparation. Your team remains responsible for the analysis and the investment decision.

Lean investment teams that need to evaluate more opportunities without dropping the standard of the analysis.

In practice that is a three-to-ten person investment team screening fifty or more opportunities a year, where the memo is still written by hand: family offices and private investment offices, independent sponsors, small and mid-market private equity, venture and private credit funds, and corporate development teams running M&A without a dedicated investment bench.

It is a poor fit at both extremes. A principal doing five deals a year has occasional pain and a general AI assistant is enough. A firm with twenty analysts and an incumbent platform has a procurement problem, not a memo problem.

The documents that normally sit in a deal folder: CIMs, financial models, management presentations, diligence reports, DDQ responses, market research, contracts, term sheets and call notes.

Before any analysis, Vurdera checks the folder and shows you which files it can process. Image-only documents with no text layer are excluded — optical character recognition is not supported yet. You know exactly what the memo is based on before it is generated, rather than discovering the gap in it afterwards.

For a first draft, the difference is small. A general assistant can produce a competent draft of an investment memo.

The difference shows up when someone asks where a number came from.

Vurdera checks each citation against the text of the source page it points to, so a claim that cannot be supported is flagged rather than quietly included. It states what the documents do not answer instead of filling the gap. It produces the same structure on every deal, so memos are comparable across a year of deal flow. It leaves an audit trail your committee can read six months later. And it does all of this under commercial terms written for confidential deal documents, rather than a consumer chat window.

Many of them focus on what is inside the documents: extraction, summarisation, search and citation. They are good at it.

Vurdera works from the deal folder and produces your firm’s investment memo. Each claim is cited to its source, missing information is flagged, and committee members can ask questions against the underlying deal documents. The process is the same on every deal.

The comparison on our home page shows what changes compared with preparing the memo by hand. If you are evaluating a specific alternative, ask and we will tell you where it is stronger.

No. Vurdera handles the document-heavy preparation. The judgment stays with your team.

Someone still has to decide what the evidence means, form a view, defend it to a committee, and own the outcome. Vurdera produces the draft that person starts from, and surfaces questions they may not have thought to ask.

Yes. Each deal gets a read-only seat for the investment committee.

A member can open the memo, ask a question about the deal in their own words and get an answer drawn only from that deal’s own documents, with the source attached.

They cannot edit the memo or the audit trail. Their questions are kept in a log, and any the documents cannot answer become items for the diligence list before the meeting.

It says so.

Ask for an IRR when there is no financial model in the data room and Vurdera will tell you the model is not there and what it would need. It does not estimate, and it does not reason its way to a number that looks plausible.

This is the behaviour we would most like you to test. It is also what makes the memo useful: gaps are identified early enough to take back to the seller before the committee meets.

It surfaces the conflict rather than picking a side. The memo cites both sources and flags the discrepancy for a human to resolve.

Contradictions between a CIM and a model, or between management’s projections and the diligence report, can be some of the most informative things in a data room. They need to be resolved by your team, so Vurdera reports them.

Yes. Vurdera produces the first draft and your team keeps control of it — edit it, challenge it, delete sections, add your own analysis, regenerate what you do not like.

Vurdera handles the preparation. The investment case belongs to your team, and the memo that reaches your committee is the one your team approved.

No, by design.

Vurdera surfaces unanswered questions, unsupported claims and the evidence relevant to the investment case. It does not issue a recommendation, score a deal, or tell you what to do with your capital.

The investment decision stays with your team.

Two checks, both mechanical.

First, every factual claim drawn from the deal materials must have a source, and each citation is checked against the text of the source page it cites — direct text and number matching, not a second model asked whether the first one was right. A claim that cannot be supported is flagged rather than included.

Second, Vurdera is built to say when the documents do not contain the answer, rather than fill the gap with an assumption.

We will not tell you it is never wrong. What we will do is show you where each factual claim came from, so that checking it takes minutes instead of a re-read.

Yes. Every factual claim in the memo links to its source document and page, and every deal carries an audit trail recording what was claimed, what supported it, and what was flagged.

The audit trail cannot be edited from within Vurdera, by your team or by us. It is there for the conversation nine months later, when someone asks what you knew at the time.

No. Your investment data is not used to train the AI models that power Vurdera.

Vurdera runs on Anthropic’s commercial API, where customer inputs and outputs are not used to train Anthropic’s models. We do not train models of our own on anything you send.

You select them on your own machine. They are transmitted over an encrypted connection, analysed, and discarded from our processing environment once the request completes.

Vurdera keeps no document repository, no file storage and no vector database. The memo itself stays on your machine.

Full detail is in our Privacy Policy. Our data processing agreement and subprocessor list are available on request, before you send anything.

No. Vurdera sits on top of the process you already have. You give it the deal materials, it produces the memo in your house format, your team reviews and edits it as they normally would, and your committee meets as it always has.

There is no data migration and no integration work. From an existing deal folder you can generate a first memo in minutes; setting up your house template takes one onboarding call.

Yes. You give us your section headings and what each section has to contain, and Vurdera drafts to that structure. We set it up with you on one onboarding call.

This matters because a memo in your own structure only needs editing. A memo in a different structure usually has to be rewritten.

Yes. Vurdera is built for repeated use across your deal flow.

Each deal is kept separate, with its own documents, memo, sources, question log and audit trail. The same structure and the same standard of sourcing apply to every opportunity, so memos can be compared across the year.

Yes. It is also how we prefer to start.

Send us something that is yours to send — one of your own past memos, a deal where you were the seller, or one where the confidentiality obligations have expired. We will run it through Vurdera and walk you through the output claim by claim, including the parts it got wrong.