Is “autonomous” doing engineering work, or marketing work?
More companies now claim an “algorithmic” or “autonomous” operating model, and the claim often drives the valuation. Financial due diligence doesn't test it. We provide the technical due diligence and architecture validation that does, for family offices, venture funds and private equity firms, and for companies that want to check their own systems before a raise.
Secular Labs reviews AI-native companies against a five-point technical framework: verification versus drift, compute-to-revenue elasticity, model portability, human-in-the-loop fallbacks, and audit-trail evidence. The framework is based on the AMEF (finance) research paper. Engagements are scoped per deal.
The five-point framework
| # | Question | What we look for |
|---|---|---|
| 01 | Deterministic verification vs. stochastic drift | Are multi-agent tasks bounded by checks that can't be skipped, or is “autonomous” mostly marketing? |
| 02 | Compute-to-revenue elasticity | Does gross margin hold if volume grows 10x, or does the cost structure quietly invert? |
| 03 | API counterparty resilience & model portability | Can core execution switch model providers without a rebuild, or is the business one price change away from margin collapse? |
| 04 | Protocol governance & human-in-the-loop fallbacks | Are high-stakes actions held for human approval, or is oversight only a policy document? |
| 05 | Empirical invariant auditing | Can the company produce an unbroken, programmatic audit trail, or is “we checked it” the only evidence? |
Who uses it
| You are | Typical use |
|---|---|
| A venture fund or family office | Pre-investment technical review of an AI-native target |
| A private equity firm | Portfolio review: which companies' AI claims hold up at scale |
| A founder before a raise | An independent check of your own architecture, so investors' questions have answers |
What you receive
A written report covering each of the five points, with findings, evidence seen, red flags and questions to put to management. It's a technical assessment, not investment advice.
The research behind it
AMEF covers how capital allocation, valuation and investment should change when coordination becomes algorithmic, including the cost of the security layer.
Frequently asked
What is AI due diligence?
A technical review of whether a company's AI systems work, scale and stay under control as described, alongside the financial and legal diligence.
How long does it take?
It depends on access and scope; we confirm a timeline per deal.
Do you sign an NDA?
Yes, before any material is shared.
Is this investment advice?
No. It's a technical assessment for your own decision process.
Can a company use this on itself?
Yes. Founders use it to check their architecture before investors do.
Discuss a specific target or portfolio.
Engagements are scoped per deal. Tell us about the company and timeline, and we'll reply with scope and fee.
Related services: AI-native operating model · AI workflow automation · AI governance & verification · Applied systems · Explainer: The verification premium