Mandate / M&A Buyers & Corporate Development
Target Screening & Data Room Auditing
Every seller document is a claim to be tested.Not ground truth.
Expose non-obvious structural risks, inflated synergy narratives, and reporting anomalies within the target's financial and operational data room before signing the Letter of Intent.
01 / The mandate
A data room is a set of management assertions.
Confidential Information Memorandums (CIM) and Virtual Data Rooms (VDR) are designed to present the target in the strongest commercially relevant light, not to provide an independent account of reality. True intelligence treats these assets as a series of management assertions waiting to be systematically disassembled.
02 / The vault
Two layers of inspected inputs.
Outside-In Company Intelligence
Publicly traceable sources only. Continuous scanning of historical corporate registries, business networks, employee movement vectors, and supplier digital footprints.
Inside-Out Private Data Room
Uploads are in testing. Confidential corporate uploads including financial models (XLSX), raw trial balances, customer concentration ledgers, quality of earnings reports (QofE), and unredacted board presentations.
03 / The evidence ledger
Nothing is synthesized without explicit grounding.
Document Evidence
Operational claims made by management, stamped with precise page, paragraph, or spreadsheet row-level locators.
Arithmetic Reconciliation
Deterministic checks on the target's financial models to isolate hidden gaps between segment revenues, gross margins, and reported EBITDA.
External Observations
Internal management claims cross-referenced against external public signals to verify operational reality.
04 / The intelligence state
Four voices. One verdict.
Inside the analysis chamber, the data room is subjected to a four-voice dialectic stress test that challenges every assumption and narrative surrounding the deal.
CREATORFormulates the target's bull-case growth thesis.ADVOCATEDefends the logic and interests of the target's management.DEVILHunts for inconsistencies, financial anomalies, hidden liabilities, and breakpoints.JUDGEWeighs the conflicting voices purely on evidence and issues the verdict.
The Epistemic Calibration Ladder
FACTConfirmed by a register, filing or authority (T1), by two independent T1 or T2 sources (T2: independent reputable media or research), or a seller-supplied (T3) figure reconciled with one of them. Seller documents are never FACT on their own; an audited statement is FACT only with an unqualified opinion and a verified register match. T4 (aggregators, reviews, hearsay, memory) is never FACT.OBSERVATIONA directly perceived signal or data point from the public or private vault.INFERENCEA logical deduction derived from observations, with its evidentiary basis explicitly recorded.ASSUMPTIONA management projection or claim that currently lacks supporting evidence.CONTRADICTIONHard anomalies where two independent data sources directly oppose each other.UNKNOWNMissing data points where no evidence is present. Where evidence is missing, Saloa says so.JUDGEMENTThe final, weighted conclusion issued by the internal Judge.
05 / The Socratic dialogue
The interrogation loop.
Instead of static reporting, the audit enforces a live research cycle. From every Contradiction and Unknown, Saloa generates targeted questions that give the buying team precise lines of inquiry for the seller's management.
Discovered friction
Stated gross margins in the PDF presentation do not align with cell formulas in row 142 of the raw financial model.
Enforced interrogation
"Explain the €5.2M variance between the reported gross margin in the Q3 deck and the underlying formula logic in the provided Excel data room."
06 / The living intelligence state
A versioned state, not a static document.
Every new document, management answer, or verified observation becomes new evidence. Saloa recalculates the state, flags which contradictions disappear and which emerge, and recalibrates the judgement while recording exactly what changed, and why.
07 / The decision contract
One explicit institutional status.
No colorful charts. No dashboard noise. A minimalist, high-contrast dossier routed into one of five formal categories.
NO_ACTIONACQUIRE_INFORMATIONMORE_EVIDENCEHUMAN_DECISIONDEEP_AUDIT_REQUIREDFICTIONAL example output · Status: DEEP_AUDIT_REQUIRED
Grounding. Revenue retention metrics are mathematically inconsistent with the customer concentration ledger. A €5.2M margin gap has been isolated in the core forecasting model and classified as a CONTRADICTION.
Prediction boundary. A prediction is issued only when sufficient historical evidence supports it. Its evidence basis, assumptions and time horizon remain visible. Otherwise, the prediction stays UNKNOWN.
FICTIONAL example prediction. If the current churn vector continues unmitigated, Q4 gross margins are projected to contract by 8.4%, breaking the target's debt covenants within 180 days post-acquisition.
08 / The single move
The investment decision remains human.
FICTIONAL next approved move · DEEP_AUDIT_REQUIRED
Open a deep audit of the customer concentration variance: request the verified billing ledgers from the seller's CFO before further diligence proceeds.
09 / Boundaries & sovereignty
Your data room stays yours.
No AI training
Saloa does not train on your data.
Tenant isolation
Row-Level Security scopes every data room to its owner. In Saloa's own verification run on 2 October 2026, 8 authenticated cross-account and 16 anonymous access attempts were all blocked.
EU infrastructure
Data stored in the EU. AI processing may involve providers outside the EU.
Deletion on command
Removed from live systems. Backups expire after [OWNER TO CONFIRM] days.
Epistemic integrity
Saloa labels what is fact, inference, assumption or unknown. Where evidence is missing, it says so.