Private equity has moved past asking whether AI works. The harder question now is why so little of it is running in production. FTI Consulting's 2026 Private Equity AI Radar, a survey of 200 senior PE decision makers at firms with at least $1B in AUM, puts numbers on the gap. For healthcare roll-ups the gap is wider still, and it is costing sponsors EBITDA every quarter.
What the Survey Found
FTI grouped portfolio companies by how deeply they use AI:
| AI adoption stage | Share of PortCos |
|---|---|
| Enterprise-level production | 7% |
| Production in specific functions | 12% |
| Production in specific use cases | 17% |
| Piloting | 21% |
| Experimenting | 16% |
| Limited usage | 15% |
| Not using AI | 12% |
Add it up and 36% of portfolio companies have AI in production in any form, while 43% are not materially deploying it at all. FTI also notes that depth varies within a single portfolio. One company is running AI across operations while its sister company is still experimenting.
The programs that do reach production are working. 95% of respondents say their AI initiatives are meeting or exceeding the business case. So the main thing holding returns back is that too few programs ever get to production.
Why Healthcare Roll-Ups Get Stuck in Pilots
A multi-location MSO or DSO has every reason to stall at the pilot stage:
- Every acquisition brings a different stack. Ten practices can mean several EHRs and PMSs, plus their own phone systems and their own habits. A pilot that works at one site has to be rebuilt for the next.
- Pilots are scoped to one location. A tool proves itself at the flagship office, and then the work of rolling it out to the other locations never gets a budget.
- Nobody owns the rollout. The operating partner expects the PortCo to drive it, the PortCo expects the vendor to, and the vendor sold a license.
- HIPAA slows everything down. Each new tool that touches PHI needs a BAA, a security review and a decision about where the data lives.
FTI's survey lists the same barriers across all industries. Time-to-value (29%), integration with legacy systems (28%) and siloed ownership (25%) all rank near the top.
What Moving to Production Looks Like
The way out is to deploy one operating system across every location at once, instead of adding one more point tool:
- Unify first. The Rollup Engine (AIP) connects the platform's EHRs, PMSs and phone systems and puts their data into one standardized model.
- Execute everywhere. The Operations Engine (AIT) runs scheduling, intake, recall, reviews and patient communication at every location on the same playbook.
- Baseline before go-live. KPIs are measured location by location before launch, so results are reported against a real starting point.
- Own the result. The system runs on the client's Enterprise Sovereign AI OS, so the data, configurations and workflows stay with the platform.
With Samara, an MSO or DSO platform is live, ROI-enabled and in production in under 30 days. In FTI's terms, that takes the platform from "piloting" straight to "enterprise-level production", the stage that only 7% of portfolio companies have reached.
Questions Sponsors Should Ask
- How many of our locations have AI in production today, and how many are only piloting it?
- Is the next add-on going to inherit our AI workflows on day one, or start over?
- Who is accountable for the rollout across the whole platform?
- Do we own the configurations and data, or does the vendor?
Next Step
See how Samara works for private equity, or book a demo to map your platform's path from pilot to enterprise production.
Frequently Asked Questions
What share of PE portfolio companies run AI in production?
According to FTI Consulting's 2026 Private Equity AI Radar, 36% of portfolio companies deploy AI in production across use cases, functions or the enterprise, and only 7% reach enterprise-scale deployment.
Why do healthcare roll-ups struggle to scale AI?
Acquired practices run different systems, pilots are scoped to single locations, ownership of the rollout is unclear and every tool that touches PHI needs its own compliance review.
How fast can a healthcare platform move AI into production?
With Samara, an MSO or DSO platform is live, ROI-enabled and in production in under 30 days.