Private equity operates on the basis of evidence.
Grant Thornton's AI Impact 2026 survey of 950 business leaders identified a pattern seen across several industries, including private equity: the industry is developing conviction about AI's potential faster than it can generate measurable results.
Across the survey, organizations are scaling up the use of AI without having the infrastructure to prove it's working. In private equity, this disconnect directly affects portfolio companies (portcos), where business value is effectively created.
Most funds have yet to build the necessary infrastructure to prove results at the level of investee companies, let alone consistently across diversified portfolios.
The governance infrastructure needed to generate measurable returns, and the ability to demonstrate that AI-backed decisions are auditable, defensible, and capable of generating provable value, is precisely where private equity has one of the biggest lags behind the average of other industries.
Private equity remains on the less favorable side
Private equity leaders demonstrate confidence in their ability to identify where AI can drive value. More than half say the strategy is the main driver of return on investment (ROI) in AI, higher than the average for other industries. Board engagement is also above average, with 80% already exploring or testing agency AI initiatives. By practically all indicators of intention, the sector is advancing rapidly.
However, the results still do not follow this movement. Only 24% report revenue growth associated with AI. In terms of maturity, 45% of private equity organizations are still in the pilot phase, 11 points above the average for all industries. In addition, only 5% say they have fully integrated AI into their operations, compared to 14% in the overall average.
Organizations that integrate AI fully are four times more likely to report revenue growth. With nearly half of the companies still in the testing phase, the private equity industry remains on the less favorable side of this equation. The difference is that, in private equity, the cost of remaining in this condition is measured in the value of the exit from investments, and not just in the annual performance of the companies.
Industry leaders say finance is the one that most needs focus, training, and incentives for their organizations to achieve their AI ambitions. The conditions found in most of the investee companies help explain this scenario. When a new company enters the portfolio, the finance function is often one of the first areas to undergo transformation, with new CFOs, new reporting structures, and new performance indicators.
What is missing is governance: the infrastructure capable of making AI results defensible, auditable, and traceable both at the level of the invested companies and the fund itself. Without this structure, finance cannot produce the evidence that investment committees, investors (LPs), and potential buyers will require to validate value creation.
More detailed analysis happens before governance is ready
Buyers, financial institutions, and investors (LPs) are already assessing companies' exposure to AI-related risks. An investee company (portco) that fails to answer these questions with documented evidence may face unnecessary impacts on its valuation. Similarly, a fund that fails to demonstrate this visibility across the portfolio tends to face the same questions on an even larger scale. Most private equity firms are not yet prepared for this.
In the private equity industry, governance means:
Clearly defined and documented accountability for AI-generated outcomes at the level of investee companies.
Portfolio-wide visibility into which AI solutions are in operation, who is responsible for them, and what evidence proves their expected performance.
Ability to present auditable and defensible evidence that AI-backed decisions can withstand the scrutiny of buyers, financiers, regulators, and investors.
Only 9% of private equity industry respondents say they are confident they would be able to demonstrate these capabilities in the next 90 days. This is the lowest percentage recorded among all the industries surveyed and an index less than half of the general average.
In addition, only 7% of private equity industry respondents have a tested AI incident response plan, with defined processes, formally designated assignees, and validation through simulations and structured exercises. Private equity firms frequently stress test financial models, simulate adverse scenarios, and execute response plans for operational disruptions. However, AI systems that operate alongside these processes still don't go through the same level of preparation.
Governance needs also vary among the investee companies in the portfolio. Organizations in the healthcare and financial services industries, for example, pose significantly higher AI-related risks than manufacturing or retail companies, due to regulatory requirements, fiduciary responsibilities, and the sensitivity of the data involved. At the same time, the pressure is mounting: 99% of private equity respondents say they are exploring, testing, scaling up, or incorporating agency AI into their operations. In many cases, however, this progress occurs without the necessary governance structure.
Deviations in models that support revenue projections, unauthorized access to data identified during due diligence processes by potential buyers, and chain failures in critical workflows are concrete examples of the risks that can arise when there is no clear definition of responsibilities or mechanisms to control and contain the behavior of AI systems.
The challenges of investee companies do not have a single solution
Companies invested by private equity funds are not advancing at the same pace with AI, nor do they face the same risks. An operating partner that applies a single approach to the entire portfolio is not managing AI effectively. Each industry, and each company within the portfolio, requires a specific strategy.
Technology
40% of tech companies surveyed say they are fully integrating AI into operations, compared to just 5% in private equity. The risk for tech-invested companies lies not in the adoption of AI, but in the inability to properly govern, measure, and document their results to withstand the scrutiny of buyers or financiers.
Manufacturing
The manufacturing industry is focusing AI on operations, where results can be captured within a typical investment cycle. However, only 7% have an AI incident response plan tested, the lowest percentage among all industries analyzed. AI is already present on the factory floor without a validated process for dealing with failures.
Financial Industry
Half of respondents from the financial industries point to governance and compliance barriers as a direct cause of AI's underperformance. Use cases have already been expanded to customer onboarding, fraud detection, and compliance processes, but many organizations have yet to clearly define who responds when something goes wrong. The question becomes whether governance evidence will be available when it is required.
Insurance
Insurers report 52% AI-driven revenue growth, one of the best results in the survey. But the governance structure that underpins these gains is still fragile: only 24 percent say they trust their controls would withstand independent review. For investee companies in the insurance industry, the risk is that the bottom line will not stand up to the level of diligence required in a transaction.
Energy
Governance barriers are the main cause of the underperformance of AI in the energy industry, being cited 11 percentage points above the average of all industries. The results of AI directly impact processes such as emissions reporting, network management, and security monitoring, which significantly raises the required control standard. Boards of directors are already asking questions about accountability that the operation is still unable to answer.
Real Estate
Real estate industry boards approve higher-than-average AI investments, but set formal governance policies at a slower pace than in other industries. In addition, a suitable approach for a construction company is unlikely to meet the needs of a real estate trust (REIT). A one-size-fits-all strategy doesn't work.
Health
The healthcare industry was not included in this survey, but it has higher AI governance risks than those seen in much of the assets in a private equity portfolio. Specific regulations related to data privacy, information security, and reimbursement models make governance failures an exposure that can directly affect the fund. Therefore, governance needs to be incorporated from day one.
Retail & Consumer Brands
The retail industry did not participate in the survey, but AI applications aimed at demand forecasting, price optimization, and inventory automation are already generating measurable returns within typical private equity investment horizons. The main challenge is in the data architecture. Gains do not expand through an omnichannel operation without a solid data foundation to support them.
The observed pattern is consistent: the gap between AI ambition and governance is real across industries, but it manifests itself in different ways. Funds that build portfolio-wide governance visibility will be better prepared to answer the questions that buyers, financiers, and investors are already asking. Those who treat AI as a single, standardized initiative for the entire portfolio may be taking risks that will only become visible when they are identified by third parties.
Three actions to generate value with AI in the invested companies
The funds that will be able to generate measurable value with AI in the invested companies will be those that build the necessary infrastructure to transform activity in AI into verifiable, measurable and defensible value.
- Develop governance visibility across the portfolio
Start by creating a documented inventory: what AI solutions are in operation at each investee company, what stage they are in, who is responsible for them, and what governance mechanisms already exist. Then, classify the risks according to the type of asset and the level of regulatory exposure. This inventory allows operating partners to confidently answer questions related to governance in any portfolio company. - Structure the financial and data foundations before scaling AI
For every investee company where AI is already in use or planned, you need to answer a fundamental question: Can the CFO prove the return on investment (ROI)?
If the answer is no, the organization must build the measurement infrastructure before scaling up the adoption of the technology. This includes consistent definitions of data, clear responsibilities over reporting, and an evidence framework that connects AI initiatives to the bottom line. - Define where AI creates value in each asset type before scaling
Map each investee company's AI opportunities by considering its operating model, margin structure, and regulatory environment. What generates value in a manufacturing company will not always be suitable for an asset in the healthcare or financial services industry.
Funds that are building specific strategies by asset class are developing sustainable competitive advantages. Those that implement AI broadly and indiscriminately, on the other hand, tend to fund experiments rather than create value.
Grant Thornton already works in this work with several private equity portfolios. The result is the construction of the infrastructure necessary to support value creation, allowing operating partners to have visibility into their investee companies and present documented evidence that AI is being governed, measured, and driving results. This is what helps protect the exit value of investments.