TLDR: Private equity has adopted AI faster than it has proven the returns. Roughly two-thirds of European funds now use AI in value creation, while around 8% report a material impact on earnings before interest, taxes, depreciation and amortisation (EBITDA) or exit value. The sponsors who turn AI into multiple will aim it at the operational levers that now drive most of the return, and they will start before close.
Adoption raced ahead of evidence
Within a single year, artificial intelligence (AI) moved from the margins of private equity to the centre of the value-creation conversation. In Alvarez & Marsal’s fifth annual European survey, 63% of funds now use AI as part of their value-creation activity, up from 41% the year before. Around 39% report deploying it across several portfolio-company functions and delivering measurable value.
The optimism runs wide. A separate May survey of 100 private equity executives reported by PitchBook found that 73% of executives expect AI to raise portfolio value over the next twelve months. The proof trails the enthusiasm. A modest 8% describe themselves as leading, meaning AI moves EBITDA or reshapes the exit story in a material way.
Where AI actually touches earnings
The common use cases cluster around analysis and the back office. Data analysis and insight generation leads at 69% of respondents, followed by operational efficiency at 60% and finance-function work at 55%. Across the firms PitchBook surveyed, the single most common application is financial planning and analysis, the most back-office function on the list.
That distinction carries the whole argument. As A&M’s Anil Kumar frames it, AI today is “an efficiency play.” BCG’s Christy Carter draws the sharper line: using a chatbot to make an analyst’s day easier and using AI to collect receivables faster and free up working capital are two very different things. The first lifts personal productivity. A receivables engine reaches the profit-and-loss statement.
The distance between the pilot and the P&L
Most of the industry sits in the experimentation phase. Cost and uncertain return rank as the top barrier, cited by 60% of funds, with data quality and availability close behind at 45%. Pilots multiply, dashboards improve, and the earnings line moves at its own pace.
The execution data tells a parallel story. Funds front-load effort, with 58% deploying value-creation resources inside the first 100 days, double the share a year earlier. Even so, 65% concede they captured under half of the value targeted in plans built over the past two years. The ambition is real. Conversion rates have room to climb.
Why this lands hardest in the mid-market
Here is the structural point worth sitting with. Operational improvement now carries the return. A&M’s analysis of European exits shows EBITDA margin improvement accounted for 51% of EBITDA growth for companies sold in 2025, up from 21.5% before 2023. The contribution from top-line growth fell from 78.5% to 49% over the same window.
When margin does the heavy lifting, AI earns its keep where it touches margin: pricing, procurement, demand forecasting, finance automation. For mid-market companies across the Switzerland, Western Europe and United States corridor, that is both the opening and the test. Smaller portfolio companies run leaner data estates and thinner technology teams, so the cost-versus-return barrier bites harder. The reward sits on the other side of it. A single well-aimed use case, a pricing engine or an automated month-end close, can move the margin of a company with revenue above 100 million euros in a way that shows up at exit.
The discipline that turns AI into multiple
The funds that pull ahead will look operational. They name the earnings lever before they buy the tool, build the data foundation the use case needs, and measure the result in basis points of margin, the unit that shows up at exit. A&M’s Bob Rajan puts the same idea plainly: the most effective use cases focus on pricing, procurement, forecasting and finance automation, where better data converts directly into margin and faster decisions.
For buyers and sellers in the mid-market, this reframes diligence. An AI roadmap reads as credible when it maps to specific cash and margin levers, each with a baseline and an owner. It reads as a pilot when it lists tools and aspirations. The same gap that separates the 8% from the rest will increasingly separate the assets that command a premium from the ones that clear at a discount. The proof will live in the P&L, and the market has started to price the difference.
References
- Alvarez & Marsal. Private equity firms turn to operational value creation as geopolitical shocks derail deal recovery (European Private Equity Value Creation Report 2026), 19 May 2026. https://www.alvarezandmarsal.com/press-release/private-equity-firms-turn-to-operational-value-creation-as-geopolitical-shocks-derail-deal-recovery
- PitchBook. PE has yet to prove its AI bets to investors, June 2026. https://pitchbook.com/news/articles/pe-has-yet-to-prove-its-ai-bets-to-investors
