Recommended Free Tools
Assess AI exposure by identifying where a company sits in the AI value chain, what specific systems it uses or sells, and whether there is credible evidence that those activities matter to its business. Then examine supplier and data dependencies, risks and controls, and the quality of the company’s disclosures. AI adoption by itself does not establish a lasting competitive advantage or an investment return.
What counts as AI exposure?
Exposure is broader than selling an AI product. A company may provide digital, physical, or financial inputs to AI; develop or integrate AI systems; or use them in its operations, products, or services. Its exposure can also come through business relationships with companies that develop or deploy systems it depends on.
Start by locating the company’s role, then identify the relevant activities and relationships. The OECD’s OECD Due Diligence Guidance for Responsible AI, published February 19, 2026, recommends understanding AI uses within an enterprise and the relationships involved in developing or deploying systems.
How to assess a company’s AI exposure
1. Map its role and specific use cases
For each material AI activity, record the business function, intended user, system or provider, data involved, intended outcome, and deployment status. Distinguish systems already in use from pilots and announcements about future plans. A broad claim that a company is “using AI” tells you little about how central it is to the business.
#1 Best Overall
2. Test the business rationale and evidence
Ask what problem the system is meant to solve and why AI is appropriate for it. The OECD guidance says investors can request a clear, concise rationale for adoption. Then ask what evidence links that use case to the claimed result: for example, a measured change in costs, revenue, service quality, or capacity. Treat forecasts and management claims as claims until the company provides support.
Determine whether any expected or realized effect is material to the investment thesis. The cited guidance does not supply a universal return metric, and adoption alone does not show that an investment will perform better.
Rank #2
3. Examine costs and dependencies
Consider the full operating picture, not just the expected benefit. Ask what implementation and ongoing costs the company bears, and whether it depends on a small number of providers for models, cloud services, compute, data, or integration. Consider whether it can switch providers and what contractual or operational constraints could make a change difficult. These questions need company-specific evidence; there is no single dependency measure that applies to every business.
4. Assess risks and governance
Review risks that fit the company’s use cases, including data provenance, privacy, bias, performance and robustness, explainability, cybersecurity, human oversight, incident response, and accountability for harm. A risk that matters for one application may be immaterial to another.
Rank #3
An IMF technical note by Xiang-Li Lim, Puja Singh, and Richard Stobo, published December 24, 2025, discusses data risks such as privacy and bias; performance issues including robustness, synthetic data, and explainability; cyber threats such as data-manipulation attacks; and broader financial-stability risks. It focuses on securities markets, not the prospects of any particular issuer. Its authors also state that their views should not be taken as necessarily representing those of the IMF, its Executive Board, or IMF management.
The OECD guidance frames due diligence as an ongoing process: identify and assess actual and potential adverse impacts, prevent or mitigate them, track results, communicate actions, and provide for or cooperate in remediation when appropriate. An investor can use that sequence to ask whether a company’s controls address the risks it has identified.
Rank #4
- Ideal for Gifting
- Ideal for a bookworm
- Compact for travelling
5. Read disclosures and follow up
Review filings and other official company disclosures for the systems involved, business purpose, material dependencies, risk ownership, controls, incidents, and measures of results. Compare what the company says about potential benefits with what it says about costs and risks. Specific, consistent explanations are more useful than broad AI claims, but disclosure alone does not verify that benefits have been achieved.
In March 2025, SEC Commissioner Caroline Crenshaw asked: “What disclosures are being made around AI uses and risk, and are they consistent and sufficient?” Her remarks at an SEC roundtable raise questions for investors; they are not a binding disclosure rule or a complete checklist.
Best Value
- Comes with secure packaging
- Easy to read text
- It can be a gift option
If public information leaves important gaps, ask management for its adoption rationale, risk assessments, mitigation plans, implementation measures, and results. The OECD guidance notes that when business relationships do not provide enough information, an enterprise may draw on existing assessments while continuing to engage for disclosure. It also gives engagement examples including bilateral dialogue with investees, requests for information or action, and considering escalation if other methods fail.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How to compare AI exposure across investments
When comparing genuine alternatives, use the same questions for each company. This framework is a practical synthesis of the OECD, SEC, and IMF material, not an official rating or standardized scoring model.
| Assessment area | What to compare |
|---|---|
| Role and use case | Where AI enters the business, which systems are in use, and how central those activities are. |
| Evidence and materiality | Whether claimed benefits are measured or realized and whether they matter to the investment thesis. |
| Dependencies | Reliance on vendors, data, compute, and integration, including switching constraints. |
| Risk and governance | Identified impacts, controls, accountability, monitoring, and remediation. |
| Disclosure quality | Specificity, consistency, and the company’s ability to answer follow-up questions. |
| Engagement capacity | Access to management and credible ways to seek more information or improvement. |
Do not turn sparse disclosures into a confident score. State what is known, what remains unclear, and how much the investment case depends on the unanswered questions.
What this assessment can—and cannot—tell you
The OECD guidance is responsible-business-conduct guidance for organizations across the AI value chain, not a securities valuation model or company rating. The IMF note concerns regulation and risk in securities markets, and Crenshaw’s remarks offer oversight questions rather than a binding standard. Together, these sources support a diligence process; they do not establish any named company’s AI exposure, current valuation, or future performance. Those judgments require current company disclosures and verified, company-specific evidence.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Clear out junk files and repair common Windows errors3Scan for outdated or missing drivers - takes under a minuteThe OECD guidance cites global annual AI venture-capital investment rising from about USD 6.4 billion in 2012 to USD 147 billion in 2024, accounting for 56% of the value of all venture-capital investment by Q3 2025. That figure describes venture-capital investment, not public-market returns, and does not show that AI adoption creates value for a particular company.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
