AI agents for financial due diligence

The next era of due diligence.

Messy seller exports in. Source-linked workpapers out: EBITDA bridge, net debt, NWC peg. Hours, not weeks.

Talk to us Meet the agents

Built for QoE boutiques · transaction advisory · ETA diligence

Today's reality

Diligence teams are stuck between two bad options.

OPTION A

Throw analysts at messy exports.

Senior people spend nights re-keying trial balances before analysis even starts.

OPTION B

Point generic AI chat at the data room.

Chat can summarize a contract. It cannot build a workpaper or defend an add-back.

Both leak margin. Teams lose 8–15 hours per engagement before analysis begins.

The three FDD workstreams

Built around the numbers that move deal value.

Every engagement comes down to three numbers in the equity bridge. Our agents are organised around them.

01Quality of Earnings

Bridge Agent

Agents draft the normalisation and pro-forma adjustments, each with a source row, rationale, and confidence level.

At an 8.0× multiple, every accepted $1 moves enterprise value by $8.

02Net Debt

Cash Agent

Agents sweep the trial balance and agreements for debt-like items the reported number misses: trapped cash, accrued interest, rent-free accruals.

Dollar-for-dollar impact on equity value.

03Net Working Capital

Cash Agent

Agents build monthly NWC, normalise distortions, and propose a defensible 12-month median peg.

Dollar-for-dollar impact on equity value.

Equity bridge · Project Alpineillustrative

Adjusted EBITDAQoE$2,423,400

EBITDA multipleclient's call8.0×

Enterprise value$19,387,200

Adjusted net debtNet debt−$4,120,000

NWC under-delivery vs pegNWC−$185,000

Equity value$15,082,200

Every line is a workstream. Every workstream lands as an Excel workpaper.

Also in the first passHistorical tradingCustomer concentrationProduct & service lineCash flow & capexForecast reviewSee the full process →

Agents as a service

Not software to learn. A team to hire.

Five agents staff the grunt work of an engagement. Exports in, source-linked Excel workpapers out. Your reviewers approve every finding.

Alpine_QoE_Workpapers.xlsx · Intake Agentrunning

B5fx=COUNTIF(Status,"unmapped")

ABC

1TB_FY22-24_final(3).xlsxTrial balance98% mapped

2gl_dump_all.csvGeneral ledger96% mapped

3bank stmts Q1-Q4.pdfBank12/12 parsed

4rev by cust FINAL.xlsxRevenue14 merged

5Unmapped accounts6 rows→ reviewer

How it works

A first pass you can defend, not a black box.

Upload the exports

Trial balance, GL, bank statements, agings. Excel, CSV, or PDF. No integrations needed.

Agents build the first pass

Bridge, proof of cash, red flags. Every number tied to a source row.

Your reviewer decides

Accept, reject, or edit every finding. Nothing ships without your judgment.

Export the workpapers

Excel workpaper pack, red-flag log, draft narrative, source index.

Not a replacement for your team. A traceable first pass, so doers become reviewers.

Security

Built for deal data under NDA.

Encrypted end to end

Engagement data is encrypted in transit and at rest.

Never trains models

Your client data is never used to train or fine-tune models.

Role-based access

Engagement-level permissions for analysts, reviewers, and partners.

Full audit trail

Every upload, finding, and reviewer decision is logged immutably.

Source lineage

Every number traces to the file, sheet, and row it came from.

Retention you control

Engagement data is deleted on your schedule, not ours.

Read our full security practices →

Questions

Asked by every partner we meet.

Quality of earnings, net debt, and net working capital, plus the supporting analyses: historical trading, customer concentration, product and service line, cash flow and capex, and forecast review.

Design partners

Run it on a real engagement.

We are onboarding a small cohort of design partners. Bring a historical engagement and measure the hours saved yourself.