Due Diligence Checklist for AI Acquisitions
The 47-item due diligence checklist every AI startup buyer needs — financial, legal, model, data and operational risk.
Cover five buckets: financial (Stripe/bank reconciliation), legal (IP, contracts, TOS), model (weights, training data, provider lock-in), data (PII, retention, deletion) and operational (SLAs, on-call, cost per request).
Financial diligence
Reconcile Stripe against bank deposits for 12 months. Match refund rates to public reviews. Confirm the MRR trend against product analytics — mismatches are the number-one hidden risk.
Model & IP diligence
Get a written statement of every third-party model used, its license, and rate limits. For fine-tuned models, verify training-data provenance — a single scraped dataset can void a deal.
Data & compliance
Confirm GDPR/CCPA deletion flows work. Audit prompt logs for stored PII. Verify SOC2 status if enterprise customers exist.
Frequently asked questions
How long should diligence take?
For sub-$500k deals, 10–14 days. Longer than that usually means the seller is stalling or the buyer is a tire-kicker.