Imagine two private-equity (PE)–owned companies both adopting AI but in different ways. Company A applies AI tools to automate administrative and accounting tasks—invoice generation, meeting scheduling, and the like. Company B goes further: it builds an agent to handle client onboarding and renewals and launches an AI-powered sales-management product for existing customers. Over time, which firm is likely to command a higher valuation?
This analysis, produced by Alfonso Pulido, Hayk Yegoryan, Joachim Bleys, and Stacey Haas with contributions from Chris Lin, Jayesh Gautam, Monica Brondholt, and Shona Sharma of McKinsey’s Private Capital and Business Building practices, finds a clear answer: companies that embed AI across operations, offerings, and new business lines show materially greater growth potential. In our sample of 471 PE-backed firms, broadly AI-embracing companies traded at a median revenue multiple roughly 130 percent higher than firms that used AI mainly opportunistically.
Data and methodology
The dataset covers 471 privately held companies that received equity or debt financing from a private-equity fund at some point. We focused on deals from 2023 onward and matching reported revenue data to capture the era of wider enterprise AI adoption. The sample was restricted to companies with revenue between $1 million and $250 million and valuation-to-revenue multiples between 5x and 300x. The companies span 30 countries and 31 industries, with a median founding year of 2015.
To make revenue-per-employee comparisons across years, we adjusted revenues using country- and year-specific inflation rates from the International Monetary Fund and backdated employee counts to the revenue year.
Firms were rated into four AI maturity levels based on an evidence-based review of company websites, regulatory filings, product documentation, and press releases available in the valuation year. The final distribution was 137 firms at level one, 213 at level two, 100 at level three, and 21 at level four.
Why this matters for private equity
With rising exit backlogs and modest returns, PE firms are increasingly prioritizing revenue growth and operational efficiency to generate portfolio returns. AI is a powerful yet still-underexploited lever. Although adoption has accelerated, many PE stakeholders are still awaiting financial outcomes: the most common AI use cases—productivity enhancements—have not consistently translated into durable revenue uplift or improved exit multiples.
Our cross-industry analysis (including software, consumer nondurables, financial services, insurance, and infrastructure) identifies four capability levels: opportunistic adoption (level one), operating-model enhancement (level two), embedding AI in products and services (level three), and business building (level four). Each higher level entails deeper technology integration, greater AI maturity, and stronger value-creation outcomes. We examined how AI adoption at these levels affects revenue efficiency and valuation multiples.
Key findings include:
- Level four companies traded at a median revenue multiple of 31x (2023–2025), the highest among the four levels.
- Companies at level four saw a median revenue-per-employee increase of $180,000—a 52 percent rise from level three.
The four levels explained
-
Level 1 — Opportunistic: AI use is minimal or limited to pilots and individual productivity tools. There is little evidence of economic impact through new products, altered business models, or reconfigured value chains.
-
Level 2 — Operating-model enhancement: AI augments workflows, speeds operations, improves decision-making, and expands output without proportional head-count growth. This can lift commercial metrics (ARPU, customer lifetime value, retention). Level-2 firms generate roughly 20 percent higher revenue per employee (inflation-adjusted) than level-1 firms, and trade at a 14x revenue multiple versus 13x for level one.
Examples: A PE-owned industrial-materials supplier deployed an AI sales agent to reengage inactive e-commerce customers, achieving a 30 percent engagement rate in the pilot. A PE-backed edtech pilot doubled average order value for an initial prospect set.
-
Level 3 — AI embedded in products/services: Firms use AI to personalize experiences, optimize pricing, improve win rates, and reduce churn. The medián revenue multiple at level three is 20x—43 percent higher than level two—indicating that markets reward AI when it changes what a company sells, not just how it operates. Developing such offerings typically requires tighter collaboration among product, data, and technology teams, and revenue gains may take longer to appear.
Example: A SaaS platform offered AI agents that reduced manual finance work, surfaced cash-flow risks, and recommended actions—delivering enterprise-level finance capabilities without additional staff. Its PE owner sold the tool within three years and realized a sizeable return.
-
Level 4 — Business building: Firms use AI to create new business lines or revenue models—data monetization, adjacent AI platforms, or AI-native service layers that complement traditional offerings. Revenue streams anchored in AI are more scalable than productivity-driven gains. At level four, median revenue per employee rises to $180,000 from $118,000 at level three.
Example: One tech company launched AI-enabled customer solutions and digital services that generated advertising, referral, and subscription revenue. Combined with legacy automation, the shift cut head count by nearly 40 percent and nearly doubled revenue per employee within two years.
Three notable patterns among level-four firms:
- Advancing from level three to level four adds approximately 11 points to the median revenue multiple (from 20x to 31x).
- Both software and nonsoftware firms generate higher multiples at each level above two, especially from level three to four; at level four, software firms traded at 33x versus 24x at level three, while nonsoftware firms traded at 22x versus 15x.
- Revenue efficiency increases sharply at level four—far beyond gains seen in earlier stages.
Organizational implications and capability gaps
Achieving level four often requires capabilities that many PE firms and portfolio companies lack. PE firms are typically structured to improve existing businesses, not build new ones from scratch. Succeeding at level four may mean treating AI initiatives as standalone businesses with dedicated teams, separate go-to-market and pricing models, distinct operating structures, and P&L accountability.
Talent scarcity remains a central constraint, despite the rise of low-code/no-code platforms. Many portfolio companies are reluctant to build permanent AI teams at early levels, creating a rent-versus-buy choice for PE firms: centralize platforms and redeploy fund-level assets and experts (rent), or embed dedicated teams in portfolio companies where proprietary data or product transformation justify ongoing investment (buy). Many leading firms adopt a hybrid approach: centralized infrastructure and expertise for common use cases, and in-house teams for product innovation.
Once the right organizational model is in place, PE firms can take three moves to scale AI across portfolios:
- Focus on the transformatory potential of levels three and four, where meaningful revenue upside exists.
- Prioritize companies with proprietary data moats and integrated product/data/engineering collaboration for level-three moves.
- When pursuing level-four initiatives, treat AI efforts as businesses: solve new customer problems, accelerate product development, and create highly personalized experiences.
Conclusion
Markets are already differentiating companies that use AI for efficiency from those that use AI to build new businesses. For PE firms, the clearest path to significant value creation lies not in isolated productivity gains but in embedding AI into product offerings, operating models, and new AI-driven businesses. As more companies develop AI capabilities, the window to capture outsized value at exit narrows; firms that move early and scale AI across their highest-potential portfolio assets are likeliest to reap the largest rewards.
Authors: Alfonso Pulido (senior partner, McKinsey Southern California); Hayk Yegoryan (partner, Copenhagen); Monica Brondholt (consultant, Copenhagen); Joachim Bleys (senior partner, Miami); Stacey Haas (senior partner, Detroit); Chris Lin (consultant, Washington, DC); Jayesh Gautam (consultant, Seattle); Shona Sharma (consultant, New York). Edited by Arshiya Khullar (senior editor, Gurugram).
