The ground has shifted for FSO (functional service outsourcing) and Unitized FSP (functional service providers) contract research organizations (CROs). 

For years, the playbook for managing these types of CRO economics was familiar. Sponsors negotiated rates while CROs managed headcount and utilization around those rates .

That playbook still holds true today in theory. However, the conditions underneath it have shifted enough that it no longer produces the results it used to. Funding is tighter, sponsors are smaller and more price-sensitive, timelines are compressing ahead of the patent cliff, and AI has moved from an experiment on the roadmap to a baseline expectation in every RFP.

None of that is unique to any one segment of the market, but these pressures land differently for FSO and unitized FSP providers than they do for full-service CROs. That’s because these businesses are built on rate cards for CRAs, monitors, and other functional resources. When sponsors squeeze rates or expect more output per unit, there's no broader program fee to absorb the hit, here the unit economics are the business.

Why every CRO suddenly feels margin pressure

Four forces are converging at once:

The difficulty of today’s market 

Every CRO leader knows the difficulties of today’s market. Those who hire ahead risk carrying an expensive staff with lower utilization over time that often erodes margins. Those who hire behind risk improperly staffing studies, costing the organization future business bids and stalling growth.

Unfortunately, there's no winning position in this trade-off because the mechanism CROs can affect, headcount, was never going to solve a cost-of-delivery problem.

Sponsors negotiate the rate, and that's their lever, especially in an FSO or unitized FSP relationship where rate cards are the currency of the deal. Cost of delivery is the one variable that's entirely the CROs. Thus, FSOs and unitized FSPs are not competing on price, instead competing on the cost of delivery.

Adding to this, the industry has already wrung efficiency out of clinical operations over the last three decades through technology, centralization, offshoring, process standardization, shared services, risk-based monitoring, and process automation. Each of those reset the cost baseline in its era.

Large-scale AI is the next structural reset, and it's the one lever that’s still largely untapped.

Where the opportunity sits and what it looks like in practice 

Clinical monitoring is where the largest opportunity lies for CROs. Monitoring accounts for roughly 25 to 30% of total trial cost, even though about 96% of clinical data never changes after its initial entry (only 2.4% of critical-data queries actually originate from manual source data verification).

Oftentimes, organizations’ most expensive personnel are the CRAs who spend much of their time confirming data that was already right. For FSO and unitized FSP models, where CRA hours are the literal unit being sold, that means margin is sitting inside every single unit billed.

AI and agentic AI have turbocharged the capabilities of the CRA. Take for example a recent deployment of Medable’s clinical trial monitoring agent with a top-five global biopharmaceutical company. The agent was deployed in Phase II and Phase III oncology, RSV, and alcohol-related liver disease programs, helping deliver measurable gains. According to the sponsor, every three CRAs produced the output of 3 to 5 CRAs, with no hire, no onboarding, and no ramp time required. Additionally, clinical monitoring tasks that once took roughly an hour took roughly 23 minutes, a 62% reduction in task time. Lastly, quality findings across the trial master file dropped by 60%.

The contract is why the narrative lands harder for FSO and unitized FSP

Under a fixed FTE or headcount-based contract, hours falling means revenue falling, so automation quietly erodes the own top line unless a CRO renegotiates the whole agreement. Under a unit- or milestone-based contract, the structure most FSO and unitized FSP arrangements already use, the CRO keeps the productivity gain. Fewer hours to produce the same unit of output means margins expand without waiting on a full commercial renegotiation.

That's the structural advantage. Many CROs commercial models are already unit-native, so when a CRA produces the output of 1.5 to 1.7 CRAs, that gain shows up directly in the economics of the unit that’s already being sold, not buried inside a program-level fee sponsor's control.

The importance of a good operating model

Before AI, revenue growth and headcount growth tracked each other closely. AI-enabled leverage breaks that link, so revenue can grow while headcount and cost stay flat. That creates room to compete above an organization’s weight, bidding on volumes and timelines that would have required proportional hiring before.

A well-designed commercial and operating model turns AI productivity into business performance across three dimensions. On the operations side, the goal is to expand capacity: keep headcount tight, and deliver faster and better with the team a CRO already has. On the finance side, the goal is to capture value: improve margins with the right contracting structure, and make sure the value AI creates is actually captured rather than given away. On the commercial side, the goal is to compete differently: focus AI investment on the quality, compliance, and execution-at-scale advantages that lets a CRO win bids on more than price alone.

The first to move will reset the market 

The first CRO to lower its cost of delivery resets the price floor for the whole market. Sponsors start expecting those lower economics, competitors are forced to respond, and industry economics reset around the new baseline. The providers that wait and watch will take the biggest hit, because by the time they respond, the floor has already moved.

For FSO and unitized FSP providers specifically, this cuts both ways faster than it does for full-service CROs. Rate cards get repriced at the next RFP cycle, not the next multi-year program renewal. Moving first buys more runway before that repricing happens.

Conclusion: The ground has once again shifted 

None of this is about racing to adopt AI for its own sake. It's about recognizing that cost of delivery, not rate, is the one variable CROs actually control, and that the commercial model is already built to reward CROs for improving it. FSO and unitized FSP providers don't need to wait for a new contract structure to capture the value AI creates. The structure is already in place, they simply need to act on it.

‍