At Mural Health, we often work at the intersection of participant experience and trial operations, so we spend a lot of time watching where the industry is moving, not just where it says it's going. Last year we made five predictions for clinical research in 2025. Now we’re taking a look at how they held up, before we share our predictions for 2026.

What 2025 made clear

2025 confirmed what’s been emerging for awhile: drug discovery is scaling faster than trial execution. AI-driven drug discovery moved from experimental to embedded, with major pharma companies now treating these partnerships as core infrastructure – and the pipeline is accelerating. But the systems that turn candidates into trial starts aren't keeping pace. WCG's 2026 Trends Report puts numbers to what sites have been feeling: 

• Phase III trials have seen a 42% increase in required procedures and a 37% rise in endpoints since 2015. 
• Protocol amendments now average 3.5 per Phase III trial, up from 2.3 a decade ago.
• 45% of sites say these challenges are impacting their ability to take on new studies. 

‍

Meanwhile, participant experience remains under-operationalized. WCG's analysis found that while many studies offer reimbursements, 95% of them don't offer compensation beyond that (despite years of industry talk about patient-centricity).

The pattern is consistent: infrastructure and policy moved in 2025, but implementation lagged.

Prediction 1: We predicted that in 2025 participant financial support would become a competitive advantage for sponsors.

We expected to see a major pharma company issue a public statement framing participant financial support as a strategic priority, not a compliance issue. The headline didn’t come, but the operational groundwork leading to it did. Our partnership with ICON deployed Mural Link, giving sites a single system for payments, tax management, and travel support, so participants aren't fronting costs or chasing reimbursements. The Harley Jacobsen Act and The Clinical Trial Modernization Act were also introduced to Congress, and The One Big Beautiful Bill raised the 1099 threshold to $2,000. Each of these is an individual, quiet change, but together they indicate a slow shift.

‍Verdict: The industry isn't saying it loudly yet, but the infrastructure is moving to make this a statement through actions.

Prediction 2: We predicted that in 2025 the clinical research industry would embrace transparent site budgeting.

We expected a greater focus on efficient site budgeting processes, with increased standardization and leveraging an AI-based financial modeling system. We were close – AI moved into the prep work and began helping with standardizing templates, processing documents, and other components of financial management. The Contract Network with Mayo Clinic, IQVIA's Clinical Trial Financial Suite, and other emerging solutions indicate that this is still an area of importance, but not a solved problem.

Verdict: Full AI autonomy didn't arrive, but some of the systems that can enable it did.

Prediction 3: We predicted that in 2025 the FDA would enforce participant experience standards.

We thought the FDA might publicly delay or reject a submission over insufficient participant experience data. That didn't happen, but the FDA did clarify what "good" looks like: PFDD Guidance 3 was finalized in November 2025, clarifying how to select and validate clinical outcome assessments that actually capture what matters to patients. Guidance on digital health technologies continued making it easier to collect participant data in lighter-touch ways.

Verdict: The FDA clarified how participant experience should be measured. They haven't enforced it yet, but the standard is taking shape.

Prediction 4: We predicted that in 2025 a sponsor would announce 30%+ timeline cuts from AI.

We predicted a major sponsor would announce an AI-driven operations overhaul cutting timelines by 30% or more. No one made that announcement, but tools to get there launched, such as Medidata's Protocol Optimization, Saama's query-generation acceleration, and Veeva's AI Agents. Each tackles a different bottleneck; combined, they could add up to meaningful time savings.

Verdict: The 30% headline is still coming, but building blocks arrived in 2025.

Prediction 5: We predicted that in 2025 regulators would approve an AI-optimized protocol.

This didn't happen, but regulators took meaningful steps: PathAI's AIM-MASH received EMA qualification in March 2025, and the FDA qualified AIM-NASH as the first AI drug development tool in December 2025. Both came with defined, bounded use cases – exactly the scaffolding that makes broader approval possible.

Verdict: Not there yet, but regulators are creating the framework that will make it possible.

‍

Across all five predictions, the operational work moved faster than the public positioning. That's often how change happens in this industry — quietly, then all at once.