AI and technical due diligence
An independent read on the technology before you commit capital: what's built, what it can do, and whether the claims hold up.
Independence means no engineering work on what we assess.
What gets assessed
Before capital moves, someone needs to look at the technology the way an engineer would, and report back in language a board can use. That's AI due diligence and technical due diligence: an independent read on what a company has built, run by someone with no stake in the outcome of the deal.
The assessment covers architecture and how it holds up under real load, security posture, and the team and process behind the product. When AI is part of the pitch, that includes AI claims validation: testing whether a model does what the deck says it does, not taking the demo's word for it.
How it runs
The assessment runs to a fixed scope, in a window agreed up front, across three phases.
-
Scope
The assessment starts with what's being evaluated and against which claims, agreed with you before work begins. It runs on data-room and repository access, architecture documentation, and time with the engineering team.
-
Assess
Architecture, codebase, and security posture get reviewed. The team is interviewed about how they build and ship. Where AI is part of the product, its claims get tested against what the model does.
-
Report
Findings land in a written assessment, a risk register, and a briefing to the buying team, on the same fixed timeline the scope was agreed to.
What you get
Three things, all handed to you directly.
Written assessment
The full technical read: architecture, security posture, team and process, and whether AI claims hold up where AI is part of the product.
Risk register
Every finding ranked by how much it should move the price or change the terms, not handed over as a generic checklist.
Findings briefing
A working session with the buying team to walk through what was found and answer questions before the decision gets made.
What this is not
Not a substitute for legal or financial diligence. This is a technical read: architecture, security, team, and AI claims. Counsel and accountants still do their part.
No ongoing engineering work on what we assess. The firm that runs the assessment doesn't join the target's engineering team afterward. That's what keeps the read independent: nothing to gain from a rosier report.
Who does the work
26 years in production software · federal delivery since 2014 (NOAA, HHS, FDA, VA, CMS, IRS)
As technical lead on federal RFP and re-compete efforts at Booz Allen Hamilton, Jay Stuart contributed to more than $50M in awarded contracts (HHS, FDA). More on his record.
Related reading: We're Shipping Code We Don't Understand — And Calling It Progress
Talk about an assessment
Tell us what you're evaluating and on what timeline.