Buy-Side Diligence · Pre-LOI Triage
AI quality of earnings triage: test the CIM before you commission the QoE.
Adjusted EBITDA is a negotiation, not a fact. askOdin runs pre-LOI triage, cross-examining the seller's claims against the financials supplied with them before you spend on accounting engagements.
Uploaded files only · Pre-LOI triage · Not a QoE; no ledger, proof of cash or peg work
RAVEN Protocol · U.S. Prov. Patent No. 63/994,876
// The Add-Back Problem
Adjusted EBITDA is a negotiation, not a fact.
By the time a deal reaches diligence, the sell side has already done the work of making the numbers persuasive. The quality-of-earnings question is not “what does management say earnings are” — it is “what does the ledger support when you strip out everything that flatters the trailing twelve months.” That gap is where purchase price is won or lost. Here is the problem with the conventional read: a probabilistic model handed the management presentation will tend to reconcile the contradictions away, because it is optimizing for a clean, confident summary. We do the opposite.
// Cross-Document Triangulation
We preserve the contradiction. We do not reconcile it away.
This is not a QoE, and it is not trying to be one. It is judgment infrastructure: it tests whether the claims in a document survive the evidence beside it. The distinction matters at the architecture level: a summarizer is rewarded for collapsing conflicting inputs into one tidy answer. The RAVEN Protocol is built to hold the conflict open.
When the narrative says one thing and the financial statements say another, that divergence is the finding. RAVEN surfaces it, ties it to both source documents, and hands it to the deal team to adjudicate. LLMs optimize for persuasion. askOdin compiles for physics.
The architectural mechanics of RAVEN's triangulation engine are protected under U.S. Provisional Patent No. 63/994,876 and are not publicly disclosed.
src: The We Company, Form S-1 (Aug 14, 2019)
narrative . . . “tech-enabled” platform, SaaS-tier multiple
financials . . 15-year lease liabilities
revenue . . . . month-to-month memberships
→ XDOC-001 — cross-document contradiction
→ Clarity Score 0/100
// Triage, Not Attestation
What the triage asks, and what it leaves to the accountants.
A deal team sees far more CIMs than it can afford to commission a QoE on. The useful question before exclusivity is narrower and cheaper: does the story hold together against its own numbers?
What the triage asks
- Does a claim in the CIM survive the financial statements supplied with it?
- Is the same figure stated the same way across the CIM, the model and the lender deck?
- Where a claim and a number disagree, which documents disagree, and where?
What it does not do
- Reconcile the general ledger or trial balance.
- Test proof of cash against bank statements.
- Set or rebuild the net-working-capital peg.
- Produce a QoE report, or anything a lender can rely on.
The engine's published record is on pitch decks and public filings — see the Terminal Audits. On CIMs it runs as a scoped pilot on your own documents (how a pilot runs). We do not compete with your QoE provider; we come before them.
// The Standard
One deterministic check, reproducible across every deal.
The same engine and the same standard on every deal. The same recorded variables under the same Framework version return the same score, and the record keeps them, so a past finding is checked from its record rather than by re-running it. That is what makes the output defensible to the IC, to the lenders, and to the next buyer in the chain.
// OBJECTION HANDLING
QoE & M&A Diligence FAQ
Quality of earnings
What is an AI quality of earnings review?
As askOdin runs it, a pre-LOI triage: before you commission the formal QoE, does the story in the CIM survive the numbers supplied with it? Is adjusted EBITDA stated the same way across the CIM, the model and the lender deck, and do the narrative claims hold against the financial statements? It is screening, not attestation. It does not reconcile the general ledger, prove cash or set the working-capital peg; the formal engagement still does that, and still produces the opinion a lender relies on.
Can AI QoE replace a Big Four engagement?
No, and it should not try. Attestation is a regulated accounting product delivered by a qualified firm, and software does not produce one. What changes is sequencing: checking whether the CIM's story holds together is cheap and belongs before the letter of intent, so the engagement confirms a thesis you already tested rather than discovering one you have already priced.
What it analyses
Can askOdin produce a quality-of-earnings analysis, or does it just summarize the CIM?
Neither. It does not produce a QoE, and it does not summarize. It cross-examines the claims in one document against the numbers in the others and returns each contradiction as a citable finding, tied to both sources. Your deal team and your QoE provider adjudicate.
Does it validate EBITDA add-backs and working-capital normalization?
Not as an accounting reconciliation. It does not test add-backs against the general ledger or rebuild the working-capital peg; that is the QoE provider's work. What it tests is whether the CIM's narrative survives the financial statements supplied with it. The engine's published record is on pitch decks and public filings (see the WeWork S-1 audit); on CIMs it runs as a scoped pilot on your own documents.
Can it flag customer concentration and revenue quality across the data room?
Where the contradiction sits between documents you supply, yes: a concentration or revenue-durability claim stated one way in the CIM and contradicted by the model is the kind of finding the engine is built to hold, tied to both sources. It does not audit contracts or cash receipts.
Fit and confidentiality
Does this work for a carve-out where financials are entangled with the parent?
Only as far as the documents supplied allow. Where the carve-out financials and the parent-level disclosures contradict each other, that surfaces as a finding with both sources cited. Untangling the carve-out itself remains the accountants' work and the deal team's call.
How is this different from a virtual data room’s AI or a RAG tool over the room?
Those retrieve and summarize within a document. askOdin triangulates across documents and preserves contradictions instead of reconciling them away. We do not compete; we consume — keep the room, add the audit.
Is our confidential deal-room data used to train models?
No. Extraction runs against an external model API under a paid-tier agreement that prohibits training use, and the deterministic engine does the analysis. Processing runs under Stateless API Orchestration and Strict Data Sovereignty; raw documents are held no longer than 30 days under a documented retention ceiling and never enter any training corpus. What persists beyond that is the derived record (the verdict and the analysis), and deletion requests reach it too. The ceiling is documented policy; until its automated enforcement ships, deletion is executed on request. The detail, control by control, is on the security page.
Test the story before you sign exclusivity.
Venture capital is the last unaudited asset class. askOdin provides the infrastructure to close the gap — starting with a scoped pilot on a CIM you are looking at now.