AI value creation is the part of a private-equity value-creation plan that depends on artificial intelligence to move EBITDA or exit value: revenue the portfolio company could not reach without it, cost it could not take out without it, or a multiple the next buyer will pay for it. Most sponsors now have that line in the IC memo. Far fewer have it in the results.
Sultan Meghji leads this practice personally. Virtova’s role is the senior, independent read for private-equity sponsors, family offices, and strategic investors: whether the AI thesis can be underwritten, whether the execution is tracking, and whether the governance underneath it will hold when a regulator or the next buyer’s diligence team opens the file.
The thesis is running ahead of the execution
This year’s survey data all points the same direction. FTI Consulting’s 2026 Private Equity Value Creation Index, a survey of 555 senior PE leaders across 14 countries, found 66% reporting AI-related benefits within twelve months, up from 34% a year earlier. Speed isn’t the problem. Only 19% of high performers exceeded their AI business case, against 5% of everyone else, and 31% described AI implementation as efficient or mostly efficient.
Grant Thornton’s 2026 AI Impact Survey is blunter. 46% of private-equity leaders say they are scaling AI across functions; 24% report revenue growth from it. The same survey found 9% very confident they could pass an AI governance audit within 90 days, against 22% across industries.
That last number is the one that matters for a sponsor with regulated assets in the book. In a bank, an insurer, a health system, or a federal contractor, an AI use case that can’t survive examination doesn’t ship, and the value-creation line that depended on it doesn’t land.
What this engagement looks like
The engagement runs in six threads. Sponsors take them one at a time or in sequence across the hold.
Investment-thesis underwriting. Senior advisory at the IC table on the AI assumptions inside a deal model: what is plausible, what is bid-protective, and what is being mispriced. Usually scoped like a pre-LOI quick-look, one to two weeks around the two or three AI questions that move the bid.
AI-native target evaluation. Diligence calibrated to the realities of AI-native companies: model dependency, data-moat durability, generative-tooling cost-curve exposure, and the gap between the engineering team’s claims and what the production system does. This thread runs inside or alongside PE technology due diligence.
Value-creation plan review. An operator-grade read on whether the AI theme in a portfolio company can be executed on the platform’s actual technology, data, regulatory exposure, and engineering organization. The output is a written view of which AI initiatives belong in the plan, which need a governance prerequisite first, and which should come out.
Post-close execution oversight. Senior accountability against the AI thread of the value-creation plan: a sequenced 90/180/365-day plan, a named owner for each initiative, and a regular read for the deal team on what is tracking. Where the portfolio company needs continuous executive coverage, this runs as a fractional Chief AI Officer arrangement.
Portfolio-company AI governance. The model inventory, accountability structure, and board reporting that let AI use cases clear compliance and reach production. In regulated portcos this is the AI governance and model risk management work, scoped to what the value-creation plan needs first.
Exit readiness. Twelve to twenty-four months before sale, an honest read on the AI story the firm will tell the next buyer, the gaps to close before the process opens, and the diligence themes the acquirer will land on.
Governance is the gating item in regulated portfolio companies
In regulated sectors the rulebook decides what ships. For bank portfolio companies, SR 26-2 replaced SR 11-7 and SR 21-8 on April 17, 2026, and it explicitly leaves generative and agentic AI outside its formal scope. Those are the systems most AI value-creation plans depend on, so the bank has to build the parallel governance discipline itself. The NIST AI RMF is the reference most U.S. firms organize that work around, and the EU AI Act reaches U.S. platforms with European exposure.
A sponsor doesn’t need to carry this detail. The sponsor needs someone at the table who does, because the sequencing follows from it: which use cases can go to production now, which need a control built first, and which carry regulatory exposure the model never priced.
Virtova’s buy-side diligence asks a target for its model inventory, the data each model touches, vendor concentration, and the regulatory exposure new ownership inherits on day one. A portfolio company heading to exit should expect the next buyer to ask the same questions.
Independent of the implementer
Implementation capacity is not the scarce thing. Sponsors can buy it from the large consulting firms or from specialist AI services shops, and most portfolio companies already have a vendor in the building. What the deal team usually lacks is a read on that work from someone who isn’t selling the build.
Virtova engagements are senior-led and small by design. Virtova does not staff the implementation program. One disclosure belongs here: Virtova LLC and Frontier Foundry share ownership, and Frontier Foundry makes secured-AI products. No Virtova recommendation is contingent on engaging Frontier Foundry. Where one of its products fits a portfolio company’s regulatory surface, it is named alongside the other vendors the company could engage, and the choice is the client’s.
Who this is for
Middle-market and upper-middle-market private-equity sponsors and their operating partners; family offices and strategic investors with material AI exposure; and the portfolio companies themselves where the sponsor has put the AI thread at the center of the investment thesis. Sectors of consistent strength match Virtova’s broader practice: U.S. financial services, healthcare and life sciences, federal contractors and adjacent industrial technology, AI-native companies, and platforms with material regulatory exposure.
When the engagement is the wrong answer
If the fund needs fifty engineers inside a portfolio company next quarter, hire an implementation firm; Virtova will say so on the discovery call. The engagement is also a poor fit where the AI theme is a marketing line with no EBITDA attached, or where the sponsor wants a scoring rubric to staple to a board deck. The work is an honest read, and sometimes the read is that the AI line should come out of the plan.
Next step
Most engagements start with a 30-minute discovery call. Bring a current deal, a portfolio-company AI question, or a value-creation plan with an AI line in it, and we will tell you what scope fits.