The 95/7 gap: AI doesn't earn itself.
Pilots are everywhere. Scale is not.
One executive described AI like a race: a few companies are running with a clear plan, a large middle group is trying to keep pace, and a third group is technically participating but not really competing.
AI does not create value because a company buys it. It creates value when leadership redesigns the business around the few use cases that actually matter. AI only works when the company changes around it — training people, changing roles, rebuilding workflows, and tying AI to measurable business outcomes. That is why companies may need to spend far more on the people side of AI than on the technology itself.
The AI premium is real. It is just being earned by 7% of companies.
1. AI lives in the operating model
AI cannot sit beside the operating model. It has to become part of how the operating model works. McKinsey's framing is the clearest: workflows have to be redesigned around AI from input through quality control, not simply automated at the edges.
For portfolio companies, that is the difference between AI-enabled and AI-native. AI-enabled means the old process runs faster. AI-native means the workflow was rebuilt assuming AI is in the loop. A back-office pilot can prove activity. A redesigned workflow can prove value.
2. The 95/7 gap
FTI's data shows the real AI gap in private equity: pilots are working, but they are not scaling. The 2026 PE AI Radar found that 95% of portfolio AI initiatives are meeting or exceeding the business case, yet only 7% of portfolio companies have reached enterprise scale.
The 7% are not winning because they have more ideas. They are winning because they have more discipline. Leading funds are pulling ahead through governance, talent integration, use-case discipline, and a willingness to scale what works instead of collecting more pilots.
3. Use-case selection beats AI spending
Gartner's research pulls the conversation away from spending and toward deployment quality: CFOs gain competitive advantage from strategic AI deployment, not AI spending levels. Meanwhile, CEOs are more likely than CFOs to see AI as a growth and profit-model lever.
That gap matters because the CFO scorecard now has a new top line: AI ROI. The CEO should set ambition, but the CFO has to prove value. The best CEOs close that gap instead of letting it slow the company down.
4. What the board now tracks
The board conversation has moved from "are we doing AI?" to "what is AI earning?" Boards are now asking CFOs for the AI unit-economic model: what each scaled use case cost, what it freed up, and what it earned.
The CFOs walking in with an AI P&L are getting a different conversation from the sponsor than the ones still saying "we are investing in AI."
5. AI is the deal thesis now
AI is no longer just an operating-model question. It is becoming part of the deal thesis, with impact already showing up in software valuations, sector rotation, and LP access to foundational AI opportunities.
Sponsors are no longer underwriting AI as a generic improvement lever. They are asking which companies can use it to compound EBITDA, which business models AI will pressure, and which sectors benefit from the infrastructure and workflow changes AI requires.
6. What good AI sponsorship looks like
Good AI sponsorship is specific. "We'll invest in AI" appears in nearly every VCP today, but the pull-aways have a clearer thesis: which two functions get an AI-native rebuild in year one, which use case is the proof point, what EBITDA contribution is expected by year three, and what people, workflow, data, and governance changes are required to get there.
Most sponsors are still in the 95% on this — which means the CxO cannot wait for perfect sponsor direction before building the operating plan. When sponsor direction is generic, build the AI thesis yourself. The board will hold you to the result either way.
7. The sectors AI is repricing
AI is rewriting the sector sort, and the capital flow is the leading indicator. Software underwriting has reset as sponsors reassess the AI premium SaaS used to carry, while capital rotates toward assets tied to AI deployment rather than AI itself — data centers, power, semiconductors, connectivity, and related infrastructure.
Your sector's AI exposure is now a board-level data point. Name whether AI is a tailwind, a threat, or a forced redesign.
The data is converging: AI is now a P&L line, not a project portfolio.
The execution gap is the alpha gap. The 7% that scale are running a different operating cadence — one workflow at a time, use cases selected with discipline, people-side change funded, and AI reported as a number the board can hold management to.
AI does not earn itself.
Is your team in the 95% — or the 7%?
Get the monthly report →