Codurance

Craft AIConsulting & AssessmentAI Technical Due Diligence

Consulting

AI Technical Due Diligence

Traditional TDD tells you whether the software is good. This one tells you whether the AI is real, defensible and economically sound.

Talk about AI Technical Due Diligence See the 8 lines

Traditional due diligence

Is the software well built?

  • Architecture
  • Quality
  • Security
  • Infra and DevOps
  • Delivery capacity

+ AI overlay

Is the AI real, defensible and sound?

  • Model strategy
  • Vendor and model lock-in
  • Data (proprietary, quality and access)
  • Inference economics
  • Model reliability
  • MLOps maturity
  • AI-specific security
  • IP and regulatory
  • AI talent depth
We measure
Critical red flags per target
Time to committee readout
% of theses with failed inference economics
Sound familiar?
  • AI-washing: a thin wrapper sold as proprietary AI.
  • A thesis killed by hidden inference economics.
  • Lock-in, regulatory exposure and key-person risk.
“The pitch says it’s proprietary AI. Is it, really?”
What we do
  • Engineering baseline: architecture, quality, security, infra/DevOps and delivery capacity.
  • An overlay of 9 AI dimensions specific to the thesis.
  • Deliverable: a report with prioritized red flags and a readout with the committee.
AI overlay · 9 dimensions
14 dimensions of analysis

The engineering baseline plus 9 AI dimensions, detailed in the panel at the top of the page.

See the 9 dimensions

How it works
  1. Code/data room and interviews: access to the target.
  2. Analysis across 14 dimensions: baseline and AI overlay.
  3. Report and readout with the committee: 2 to 4 weeks per target.

Start with a due diligence of 2 to 4 weeks per target.

Scope and timeline agreed before you sign. You see the result of each step before deciding on the next.

Craft AI

The other seven ways in

AI Diagnosis & StrategyPlatform, data and team ready to run agents safely.AI First EngineeringCodurance's AI-first process inside your engineering.AI Governance ConsultA governance layer between application and model, from inventory to audit.AI ModernizationLegacy modernized in months, not years, with no big bang.AI App DevelopmentFrom idea to system in production, on an AI-first pipeline.AI Test AutomationA complete test safety net, from unit to E2E, without stopping the team.AI Feature DeliveryAn AI-first squad inside your product, multiplying throughput.