Technology and AI due diligence

What you get

A deal position, not a fifty-page appendix: quantified findings with explicit assumptions and confidence levels, ordered by what they mean for the transaction - price implications, conditions to close, or reasons to walk. I have sat on every side of this table: interim executive, hands-on architect, contractor inside scale-ups and founder of an exited B2B SaaS, so the findings read the way a deal team needs them, and stand up when engineering pushes back.

Scope

• Architecture and codebase review: scalability, technical debt, build-vs-buy decisions the target got right or wrong
• Cloud spend baseline: what the platform actually costs to run, and what it will cost at the growth case
• Team and delivery process: key-person risk, hiring plan credibility, whether the roadmap is achievable
• Security posture: identity, access control, pipeline hygiene and the gap between policy and practice
• AI due diligence: whether the AI capability is defensible or a thin layer over public model APIs

The AI question

AI due diligence is now the part of tech DD where valuations move, and it needs someone who has shipped the technology, not just read about it. I took agentic AI into production at a regulated derivatives trading venue, so when I assess what is model versus prompt, how the product behaves off the demo path, and whether training data and licensing hold up, it comes from practice. The method is written up in the AI due diligence checklist and Is the AI real, or a wrapper?

Sell-side too

If you are preparing for a buyer's tech DD rather than running one, the same review flips: find what their advisers will find, fix what can be fixed in the window, and frame honestly what cannot. Two to five days, well before the data room opens.

Pricing and timeline