Due diligence, done with evidence.
Upload a data room. Get a completeness assessment, a cited due diligence report, and a quality-of-earnings bridge where every adjustment carries the transactions behind it — reviewed and signed by your team, not by a model.
- 146
- checklist items tracked
- 100%
- of findings cited to source
- RLS
- enforced deal isolation
- Owner compensation above market+412,000
- Legal settlement, non-recurring+185,000
- Deferred maintenance run-rate−96,000
Owner compensation above market
Cited- Evidence
- 3 GL transactions, FY2025
- Source
- General Ledger 2025, p. 14 — Schedule 12
- Benchmark
- 2024 industry compensation survey
How it works
Upload the data room
Drop in the files as they arrive — financials, contracts, cap tables, leases. Processing runs in the background.
Review what's flagged
Deducto maps documents to a 146-item checklist and surfaces gaps, extractions, and a first-pass QoE bridge.
Sign off as a team
Every output is reviewed and approved by your team before it counts — the model drafts, people decide.
What you’ll need
Send whatever you have — the full 146-item checklist tracks the rest. These eight are what the analysis actually runs on, and the general ledger alone drives the entire quality-of-earnings side.
- General ledger detail (24 months)
- Trial balance
- Bank statements or a deposit schedule
- Financial statements (balance sheet and P&L)
- Customer contracts (last two years)
- Revenue by customer (last two to three years)
- Cap table, stock ledger and option grants
- Related-party and affiliate agreements
- Guarantees, contingent liabilities and litigation schedule
- IP registrations, and employee and contractor agreements
- Loan agreements, leases and insurance policies
Built for the diligence you can defend
Completeness assessment
Every data room is checked against a 146-item due diligence checklist, so you see what's missing before it becomes a surprise in diligence.
Cited findings, not black-box answers
Every figure and finding links back to the exact document and page it came from — reviewers can verify a claim in one click instead of taking the model's word for it.
Quality-of-earnings bridge
Adjustments to EBITDA are tracked against their supporting transactions, building a bridge your team can defend, not just a number.
It says when it cannot tell
A figure built on evidence the engine could not read is reported as unreliable rather than quietly built from the rest. Absent, illegible and not-applicable are different answers here, and none of them is zero.
Security
Deal isolation in the database
Postgres row-level security scopes every query to the deals you are on, enforced by the database rather than by application code that has to remember. The API role owns no tables and can be given no exception.
Every figure traces to a document
Findings cite the page they came from and adjustments cite the transactions beneath them. Where the evidence is weaker, the product says so rather than presenting the number as equally sound.
An append-only audit log
Who did what, and when, recorded on a table the application can write to and cannot rewrite. It outlives the deal it describes, so a deleted deal does not erase its own history.
The model drafts, people decide
Nothing an analyzer proposes counts until a reviewer accepts it. An adjustment sits as a proposal against its evidence until someone with a name agrees, and rejecting one is recorded too.