How AI-ready is your portfolio?

Almost every company now uses AI somewhere. In McKinsey's 2026 survey, nearly nine in ten respondents said their organization uses AI regularly, but only 37% reported any effect on operating profit. Using AI and benefiting from it are different things.

For investors, that gap cuts both ways. Companies that turn AI into better products or leaner operations can pull ahead quickly. Companies whose product can now be rebuilt cheaply with AI may find their position weaker than their last valuation suggests. Our article "Is software IP dead in the age of AI?" looks at that second risk in detail.

A consistent, company-by-company view lets a fund or accelerator see both at once, and decide where time and capital will do the most good.

These questions are deliberately simple. A founder should be able to answer each one in a few minutes, with evidence. If they cannot, that is itself a finding.

The strongest companies use AI on both sides: to improve what customers buy and to change how the business runs. BCG found the companies getting real value from AI focus on fewer opportunities, scale more of them, and get most of their value from core business functions rather than side projects.

McKinsey's 2026 data shows the same pattern. AI high performers are about three times as likely to have fundamentally redesigned workflows, and about twice as likely to have leaders who visibly support the work and a defined way to measure impact.

For a portfolio company, a simple test is whether AI shows up in the metrics the board already reviews, such as gross margin, sales cycle, support cost per customer, or release frequency. If AI is only visible in the product roadmap, it is still a plan.

When anyone can access powerful AI models, the model is rarely the advantage. What matters is what the company feeds it and where it sits in the customer's work. Proprietary data, feedback that improves the product with use, and deep integration into customer workflows are all harder to replicate than features.

Investors should test data claims rather than accept them. Where did the data come from? Does the company have the right to use it this way? Does more customer usage actually make the product better? Our article "AI due diligence starts with the data" sets out how to trace those claims back to evidence.

Pressure to look like an AI company is real, and some companies give in to it. Regulators have noticed. In March 2024 the US Securities and Exchange Commission settled charges against two investment advisers, Delphia and Global Predictions, for making false and misleading statements about their use of AI. They paid $400,000 in total. The SEC called this "AI washing" and said it hurts investors.

In September 2024 the US Federal Trade Commission announced five actions against companies using AI claims to mislead customers. One, DoNotPay, marketed itself as the "world's first robot lawyer" but had not tested whether its output matched the quality of a human lawyer. It agreed to pay $193,000. The FTC's chair said there is no AI exemption from the laws on the books.

For a fund, overstated AI claims are a valuation risk, a legal risk, and a reputational risk. Checking that a company's AI story matches what it has actually built protects both the company and its investors.

Once each company is assessed on the same four questions, sort them by two things: how much value AI could create or destroy for that business, and how ready the company is to act. This turns a long list of companies into a short, prioritised support plan.

Keep it light and consistent. A short questionnaire for founders, followed by a one-hour conversation with the technical lead, is usually enough to place a company in a segment. Deeper technical review can then be focused on the companies where it matters.

Use a shared vocabulary for risk so findings can be compared across companies. The NIST AI Risk Management Framework is a useful, widely recognised reference. For companies preparing to raise or sell, our AI technical due diligence scorecard sets out what a serious buyer or investor will test.

To check readiness inside each company in more depth, use our AI readiness assessment framework. It scores six areas, from strategy to the capacity to deliver change, and works for start-ups as well as larger businesses.

At the end of a review, a fund or accelerator should have a readiness baseline for every company, a clear view of which companies are exposed and which are positioned to win, and a prioritised plan for founder support. Repeat it each year, and before major fundraising rounds.

The goal is not to make every company an AI company. It is to make sure every company has an honest answer to how AI changes its market, and that the portfolio's support goes where it will make the biggest difference.