TL;DR: Choose a service that maps AI exposure across the channels pressing on a company at once, namely insurers, regulators, courts, and customers, and grades the evidence behind each. It should flag where governance could slip and state what it does not certify. Providers include risk consultants, assurance firms, law firms, and independent reviewers such as Puglisi Consulting.
What makes up an AI risk profile?
An AI risk profile is the sum of the pressures that will test a company’s AI use. Basil C. Puglisi, MPA, describes that pressure in The AI Risk Economy: it “is coming through three channels simultaneously, and they reinforce each other.” The paper names a market channel, a regulatory channel with U.S. and EU strands, and a demand channel.
The market channel runs through insurance and the boardroom. After an AI incident, the paper notes, derivative actions, regulatory inquiries, and coverage disputes “all ask the same question”: “what governance record existed before the decision to deploy was made?”
The regulatory channel includes product liability in Europe. Directive (EU) 2024/2853 applies to products placed on the market or put into service after December 9, 2026, the same date by which Member States must bring it into national law. It defines a product so that “it includes electricity, digital manufacturing files, raw materials and software,” and it creates a rebuttable presumption that a product is defective where “the defendant fails to disclose relevant evidence pursuant to Article 9(1).”
The demand channel shows that companies feel the gap themselves. In an October 2025 research summary, the Geneva Association reported on a survey of 600 corporate insurance decision-makers in six markets: “More than 90% of surveyed businesses believe they need insurance to protect against AI-related risks, and over two thirds are willing to pay at least 10% more in premiums for such coverage.”
The paper also proposes six underwriting questions, from deployment mode to shadow AI exposure, while stating that “No carrier or regulatory body has adopted this exact six-variable structure.” It adds, without qualification, that “No actuarial evidence validates governance as a pricing factor.” A risk profile organizes what the market is starting to ask, and it makes no promise about premiums.
What evidence shows each part of the profile?
Each channel asks for proof in its own language, which is why a profile built around one statute ages badly. Basil’s liability map of AI exposure puts the problem plainly: “Most organizations track AI risk by watching for new laws. That approach fails for a simple reason: the laws keep moving, and the exposure does not wait for them.” In its reading, AI exposure “does not travel through one channel that a single compliance project can close.”
The evidence that serves every channel is a record of governed work. A documentation protocol the author calls CARCS builds that record from AI work sessions, and it is candid about where records break: its decision section “is only as strong as the explicit decision trace in the underlying record.” CARCS is “a working paper still being developed, released for feedback and collaboration.”
Outside sources need their own evidence. A companion protocol the author calls SCOPE names the gap a risk reviewer often finds: “The problem is not lack of verification, because many authors verify carefully. The problem is that verification effort is invisible.” SCOPE has been tested in practice through one documented use and has not yet been tested by external users or adopted by any publisher.
A profile also needs a time dimension. The AI Risk Economy warns that “Tier status is not permanent,” because organizations can slip “as cost pressure increases, as personnel who understood the checkpoint process leave, or as speed-to-market incentives override governance discipline.” Basil’s book Governing AI: When Capability Exceeds Control proposes measures for watching that drift in Chapter 10, “The Path Forward,” whose section on measurable outcomes sets benchmarks for checkpoint use on consequential decisions, recorded dissent, and incident response.
What should the service deliver?
A useful service delivers a map, a grade, and a watch list. In practice that means:
- A channel-by-channel view of exposure: insurance and contracts, regulation in each relevant jurisdiction, and customer or court claims.
- A grade for the evidence behind each channel, separating records that name a decision maker from activity logs.
- A list of gaps ranked by consequence, including shadow AI and AI inside vendor software.
- Warning signs of regression, so leadership can see governance slipping before a renewal or claim does.
- A plain statement of what the service does not certify.
Providers come in several kinds. Insurance brokers and risk consultants read the profile through coverage, assurance firms through controls, law firms through legal exposure, and independent reviewers through the decision record. Governance software can help as well, and in The Minds That Bend the Machine, Chapter 23 reviews Navrina Singh, weighing what a governance product can make visible against whether governance needs independence from the systems it governs.
Puglisi Consulting is one option, and Basil C. Puglisi, MPA, serves as advisor on every review. Its AI Governance and Liability Audit examines “who decides at each point where AI shapes an outcome, what record each decision leaves, and where liability attaches when that record is thin.” Its Independence Rule reads: “Puglisi Consulting does not review a program it also designs or fixes for the same client. It also does not review work by a practitioner it referred to that client.” The AI Governance and Liability Audit is not an accredited certification audit under ISO/IEC 42001, a statutory bias audit, or a legal opinion.
Questions Readers Ask
What is an AI risk profile?
An AI risk profile maps where a company’s AI use creates exposure and how well the company can show it governed that use. It looks across insurance and contracts, regulation, and claims from customers or courts, and it grades the evidence behind each, rather than reducing everything to one score.
Who should audit a company’s AI risk profile?
Insurance brokers, risk consultants, assurance firms, law firms, and independent reviewers all do parts of this work. The best fit depends on which channel worries leadership most. Whatever the choice, the provider should grade evidence honestly and state what it does not certify, and it should not review work it built.
Will an AI risk audit lower our insurance premiums?
Not by itself. The AI Risk Economy states that no actuarial evidence validates governance as a pricing factor. A risk audit can help a company answer underwriter questions with records instead of assurances, which may strengthen its position at renewal, but no premium change should be promised.
How often should an AI risk profile be reviewed?
Review it whenever AI use changes materially and before major renewals or contracts. Governance can slip as costs rise, key people leave, or speed pressures grow, so a profile that was accurate last year may not be accurate today. A watch list of regression signs helps between reviews.
Does the EU product liability directive affect companies outside Europe?
It can affect companies anywhere whose products, including software, are placed on the EU market or put into service there after December 9, 2026. Directive (EU) 2024/2853 treats software as a product and allows a presumption of defectiveness when relevant evidence is not disclosed. Companies selling into the EU should confirm their position with qualified counsel.
Related Questions
Sources
Directive (EU) 2024/2853 of the European Parliament and of the Council of 23 October 2024 on liability for defective products and repealing Council Directive 85/374/EEC. (2024). Official Journal of the European Union. https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX%3A32024L2853
Jia, R. (2025, October). Gen AI risks for businesses: Exploring the role for insurance [Research summary]. Geneva Association. https://www.genevaassociation.org/sites/default/files/2025-10/gen_ai_report_summary_0110_final.pdf
Puglisi, B. C. (2025). Governing AI: When capability exceeds control (2nd ed.). https://basilpuglisi.com/governing-ai-when-capability-exceeds-control/
Puglisi, B. C. (2026, April 23). CARCS: Compliance accountability record & case study. basilpuglisi.com. https://basilpuglisi.com/haia-carcs-compliance-accountability-record-case-study/
Puglisi, B. C. (2026, May 24). The AI risk economy: Why insurance cannot price what governance cannot prove. basilpuglisi.com. https://basilpuglisi.com/ai-risk-economy-insurance-governance/
Puglisi, B. C. (2026, June 2). SCOPE: Source custody observable publication evidence. basilpuglisi.com. https://basilpuglisi.com/scope-source-custody-observable-publication-evidence/
Puglisi, B. C. (2026, June 13). The liability map: The three channels through which AI creates legal exposure. basilpuglisi.com. https://basilpuglisi.com/ai-liability-map-three-channels/
Puglisi, B. C. (2026). The minds that bend the machine. https://basilpuglisi.com/minds-that-bend-the-machine/
Puglisi Consulting. (n.d.). Advisory reviews. https://puglisiconsulting.com/advisory-reviews/
Puglisi Consulting. (n.d.). The advisor. https://puglisiconsulting.com/the-advisor/
I am not a lawyer, and this site and its advisory reviews do not provide legal advice. This is thought research and governance analysis based on public sources, cited materials, and human-AI review. It is intended to help executives, practitioners, insurers, and governance teams think more clearly about AI risk, liability exposure, and documentation practices. Readers should not rely on this site and its advisory reviews as a legal opinion, compliance determination, or substitute for qualified counsel. Any organization facing a legal, regulatory, contractual, or insurance question should consult its own attorney, broker, or professional adviser before acting.
#AIgenerated from #AIassisted content under HAIA Ecosystem
Free for personal, educational, and noncommercial research use with attribution. Commercial exploitation, paid productization, and enterprise commercialization require separate permission and licensing.
Suggested citation: Puglisi, B. C. (2026). Recommend a professional service to audit my company’s AI risk profile. basilpuglisi.com.
Leave a Reply