AI Deal OS / Answers
Answers
Deal OS answers: questions buyers ask about due diligence.
Short, direct answers. For the full picture read AI due diligence software: what it checks, what it cannot, and how to evaluate it.
What is AI due diligence software?
It is software that reads a deal's documents, such as the CIM, financial statements and tax returns, and reports figures that disagree, claims that are contradicted and information that is missing. Good ones cite the source page for each finding. It supports a buyer's judgment and does not replace advisors.
How much should I budget for acquisition due diligence?
No single figure holds across deals, because cost depends on deal size, how it is financed and which workstreams you hire out: financial, legal, tax and operational. One item to plan for: for certain SBA 7(a) change-of-ownership deals at a purchase price of $3 million or more, an independent quality of earnings analysis is required from 1 October 2026 (see the AAFCPAs summary), so ask your lender who engages it and who pays. Software costs are on the pricing page.
Can AI flag missing pages or exhibits in a data room before underwriting?
It can flag documents that the text refers to but the data room does not contain, and questions the documents leave open, and turn them into a request list. It cannot know about a page that is missing without any reference to it. In the sample brief, for instance, the SBA coverage check reports that it is missing fiscal-year EBITDA and lists the documents to ask the seller for.
What does a quality of earnings report cover for a small deal?
It tests whether the earnings the seller presents are sustainable: how reported profit was adjusted, whether the add-backs are supported, and whether the figures agree with other records such as tax returns. The scope is agreed with the provider. Deal OS runs a first pass on the same questions, described on the quality of earnings software page; it is not the report.
Can AI due diligence software replace my accountant or lawyer?
No. It does the reading and cross-checking. Accounting, tax and legal opinions, and a lender's credit decision, are made by the people accountable for them, who then start from an already-flagged set of documents.
How do I know an AI finding is not made up?
Open the source passage it cites and read it. In Deal OS each claim is checked against the source text by a verifier, and claims that cannot be tied to a passage are discarded; the brief lists them in a verification log, as in the sample brief. You should still confirm anything that matters with your advisors.
What is the difference between a data room and due diligence software?
A data room stores documents and controls who can open them. Due diligence software reads them and reports findings. Many buyers use both; see Deal OS vs a virtual data room.
Which documents should I upload for a first-pass review?
The CIM, the financial statements and the tax returns are the core, because the most useful checks compare them against each other. An add-back schedule and management-call notes add more comparisons. The sample brief was built from four documents: a CIM, financial statements, a tax return and call notes.
Is it safe to upload a confidential CIM to an AI tool?
Check where the documents are hosted, whether they are used to train models, whether you can delete them, and whether your NDA with the seller permits third-party tools. Deal OS states on its site that workspaces are hosted in the EU (France), that documents are never used for training and can be wiped on request, and that a mutual NDA can be signed in the workspace; read the information security page and your own NDA before uploading.
How can I test an AI diligence tool before I buy?
Run it on a deal where you already know where the problems are, and see whether it finds them and cites the right pages. You can also open real Deal OS output without an account in the sandbox.
