AI & Tech Governance Compliance
Anweshna reads product documentation, vendor contracts, and compliance filings for the AI-specific governance, dependency, and supply-chain risks that have become a standard diligence item in tech-enabled deals.
What it screens for
- EU AI Act classification — whether a system is classified prohibited or high-risk, and what use case triggers it.
- Foundation-model dependency — a named vendor with no fallback, contract exit terms, or history of unilateral access changes.
- Training-data provenance and data-moat strength — undocumented or unlicensed sourcing, unclear chain of custody, and whether a proprietary dataset could be replicated by competitors within 12 months.
- AI-talent concentration — capability held by a departure-vulnerable individual or team with undocumented model-training knowledge.
- AI supply-chain and security exposure — a missing AI bill of materials, undisclosed fourth-party/subprocessor AI models, model poisoning or backdoor risk, and open-source/copyleft contamination in AI-generated code.
- Algorithmic discrimination findings and conformity gaps — the affected population and nature of the finding, or a missing conformity assessment/CE marking.
Why it's a blocking category
AI & Tech Governance carries a 7% weight and is one of seven categories that can independently flag a deal for mandatory review — any document scoring 70 or above here halts the deal until a human reviews it. Model/vendor dependency, an unreplicable data moat, and key AI-talent concentration are now routinely cited factors in real deal walk-aways and valuation discounts, and a confirmed finding like training-data poisoning or a backdoored pretrained model is treated as terminal, on par with a confirmed criminal or regulatory finding elsewhere in the rubric.
If a document discloses nothing at all in this category, that absence itself becomes a flagged finding rather than being read as "clean."
AI & Tech Governance Compliance FAQ
What does Anweshna screen for in AI & Tech Governance Compliance?
Anweshna screens deal documents for EU AI Act classification, foundation-model dependency, and training-data provenance and data-moat strength, among other signals. See the full list of signals above.
Is AI & Tech Governance Compliance a blocking risk category?
Yes — AI & Tech Governance Compliance carries a 7% weight in Anweshna's composite score and is one of the categories that can independently flag a deal for mandatory review: any document scoring 70 or above here halts the deal until a human reviews it.