AI and Private Equity: The New EBITDA Improvement Opportunity

Private equity has always made money by doing the unglamorous things well. AI may be the most powerful unglamorous tool the industry has been given in twenty years.
The classic private equity value creation playbook is well known: operational improvements, procurement savings, pricing discipline, professionalising management, add-on acquisitions and, of course, financial engineering. With interest rates now much higher than they were during the last decade, the financial engineering part has become far less effective. Leverage is expensive. Multiple expansion cannot be taken for granted.
That puts operational improvement back at the centre of the model. And this is where AI becomes interesting.
From theory to tangible levers
In 2024 the use cases have become concrete. In February, Klarna reported that its AI assistant handled 2.3 million customer conversations in its first month, doing the equivalent work of around 700 full time agents, and estimated a profit improvement of about $40 million for the year. Similar results, at smaller scale, are appearing in mid sized companies across customer service, finance, procurement, software engineering and sales.
For a PE fund, the logic is compelling. A typical portfolio company with €200 million of revenue and €30 million of EBITDA might have tens of millions in costs related to administrative processes, customer support, back office functions and IT. A realistic AI programme that improves productivity in a few of these areas by 15 to 30 percent can add several points of EBITDA margin. At typical exit multiples, that can translate into a substantial increase in equity value.
Why most programmes underdeliver
But here is the uncomfortable truth. Most AI initiatives I see in portfolio companies fail to deliver their promised value, not because the technology does not work, but because the implementation is treated as a software purchase.
In my experience as an operator, the companies that capture the value do four things differently. They start from a P&L line, not from a technology. They redesign the process around AI rather than adding AI on top of an unchanged process. They assign clear ownership for the outcome to a business leader, not to IT. And they measure value in hard numbers from the beginning, with a baseline, a target and a timeline.
The companies that fail typically launch dozens of pilots, celebrate the demos, and discover a year later that nothing has changed in their cost base.
The role of the fund
This creates a specific role for the PE fund itself. Most portfolio companies, especially in the mid market, do not have the capabilities to run AI transformations on their own. They lack data engineering talent, they lack experience in process redesign, and their management teams are focused on hitting the budget.
The funds that will lead in this area will build central capabilities: an operating team that can diagnose AI opportunities across the portfolio, standard playbooks for common use cases, preferred vendor relationships, and the governance to track value delivery. In effect, AI becomes part of the 100 day plan.
A caution on headcount narratives
It is tempting to translate AI into headcount reduction. In some processes, that will be the outcome. But the more sustainable value often comes from capacity: serving more customers with the same team, responding faster, reducing errors, improving pricing decisions. A narrative built purely on cost cutting can damage morale and customer experience, and buyers at exit will look through it.
Investment question: Could AI become one of the most powerful drivers of value creation in Private Equity portfolios over the next five years?


