Methodology
AI does the depth. I do the judgement.
A look behind the scenes at how a typical diligence engagement is run, what the AI agents actually do, and why the combination produces sharper findings than either approach alone.
The engagement
Five business days, day by day.
Day 1 — Scope and setup
A 60–90 minute kickoff call with the company’s leadership. Read-only access is granted to repositories, infrastructure, and product documentation. AI agents are configured against the specific technology stack.
Day 2 — Technology scan
AI agents run exhaustive scans across the codebase: code quality, security, architecture patterns, dependency health, infrastructure configuration. The output is thousands of data points, ranked by significance.
Day 3 — Product and commercial review
Documented materials are reviewed. Cohort and usage data is analysed. Structured interviews with product, engineering, and commercial leadership add context that documents cannot provide.
Day 4 — Synthesis
All findings are reviewed, prioritised, and contextualised. Each is categorised under the RAG framework. The narrative is built: what matters, why it matters, and what to do about it.
Day 5 — Delivery
The full report is delivered. A 60-minute live walkthrough covers every finding, the business impact, and the recommended remediation plan. For investor engagements, an executive summary is provided for investment-committee use.
The honest answer
What the AI actually does.
AI agents read the code, the infrastructure, and the documentation faster and more thoroughly than any human team could. They flag patterns. They surface anomalies. They identify what looks like risk.
But AI doesn’t know which findings matter to your business. It doesn’t know which architecture decisions are reasonable for your stage and which are red flags. It doesn’t know how to write a recommendation a founder can actually execute. That’s the judgement layer. That’s why human review is essential, not a nice-to-have.
Speed
Five business days, not five weeks. Investors don’t have weeks. Founders don’t either.
Coverage
Every line of code, not a sample. AI doesn’t skim.
Cost
A fraction of Big 4 consultancy fees. Without the polish-over-substance trade-off.
Operator judgement
Sixty engineers building production systems daily. The findings come from someone who ships, not someone who reviews.
Want to see the methodology in practice?
Download a full sample report — same structure, same depth, same format as a real engagement.