AI Diligence Is Not a Feature Checklist
Our Perspective
These insights are synthesized from our deep experience executing M&A technical diligence and optimizing enterprise architecture. They reflect our ground-truth perspective on what investors must prioritize to separate AI hype from defensible, structural value.
AI diligence is not a feature checklist. It is not enough to ask whether a company has a chatbot, a copilot or agent, a roadmap slide, or a few generative AI experiments. For investors, AI touches deeper questions: product architecture, engineering productivity, implementation timelines, gross margin, pricing strategy, data governance, vendor dependency, security posture, competitive risk, and exit narrative.
The wrong diligence question is whether this company uses AI. A company may use AI superficially, expensively, insecurely, or without differentiation. Another company may have limited customer-facing AI but significant opportunity to use AI internally to improve engineering, support, implementation, or operations.
The better question is how AI changes the value, risk, architecture, and economics of the business. AI can strengthen a company’s position if it improves product value, increases retention, reduces support cost, accelerates implementation, improves engineering productivity, or creates differentiated insights from proprietary data. It can weaken a company if it erodes margin, creates vendor dependency, exposes sensitive data, reduces switching costs, or makes the product easier to replicate.
A practical diligence lens covers capability, architecture, defensibility, economics, and execution. Capability asks where AI lives, how it is used, and what value it unlocks. Architecture asks how it is implemented, governed, and monitored. Defensibility asks whether it strengthens or weakens the moat. Economics asks how it affects margin, pricing, services, and retention. Execution asks whether the team has governance and operating capability.
Red flags include claims without usage data, generic features with limited differentiation, single-provider dependency, no inference cost model, unclear data governance, no output monitoring, lack of human review in high-risk workflows, and roadmap-heavy narratives with limited execution.
Green flags include AI tied to measurable outcomes, real adoption data, clear model strategy, cost-aware architecture, strong data governance, provider-switching capability, human review, measurable productivity gains, differentiated domain data, and a credible roadmap linked to value creation.
AI diligence should inform post-close planning. Potential opportunities include engineering acceleration, technical debt reduction, support automation, implementation compression, premium AI modules, data strategy, buyer-facing narrative, and margin-risk reduction.
Apply this thinking to your portfolio.
Thought Source helps investors and operators assess AI architecture, defensibility, and value creation.