By Andrew Reina, Chief Revenue Officer
Clinical research is at an inflection point. Sponsors and CROs are being asked to deliver trials that start faster, enrol more predictably, and generate data that stands up to increasing regulatory and commercial scrutiny. In that environment, long-standing approaches to site selection, which are often shaped by reputation, relationships, and historical familiarity, are starting to show their limits.
Experience still matters. Relationships still matter. But they are no longer sufficient on their own.
Sponsors are increasingly asking for predictability. What matters is not just speed, but confidence that a site will perform as expected against the specifics of a protocol. That shift in expectation is what led Velocity Clinical Research to develop an AI-powered feasibility and site intelligence tool, to bring greater rigor, transparency, and consistency to how site selection decisions are made.
From feasibility-as-opinion to feasibility-as-evidence
At its core, the tool takes a clinical protocol and evaluates it against a broad range of site-level data drawn from across Velocity’s global site network. This includes investigator experience, historic enrolment performance, operational delivery metrics, prior protocol execution, and therapeutic expertise. Using AI, the system assesses how closely each site’s real-world performance aligns with the demands of a given study.
The output is a ranked view of sites most likely to succeed, accompanied by the underlying performance signals that inform those recommendations. Rather than asking sponsors to rely on intuition or precedent, the tool makes the reasoning behind feasibility decisions explicit, defensible, and data-driven.
Site selection has an outsized impact on timelines and cost. By quantifying performance in a consistent way, AI enables feasibility to become more repeatable and less subjective. For an industry increasingly focused on data integrity and execution certainty, this tool represents an important evolution.
Predictability as a driver of data quality
Predictability is foundational to data quality. Sites that are well matched to a protocol are more likely to recruit on time, adhere to study requirements, and deliver consistent, reliable data. When that alignment is missing, the consequences often appear downstream in the form of costly delays, protocol amendments, or uneven data sets.
Velocity’s feasibility creates a confidence index: a quantitative measure of how well a site’s historic performance patterns align with the specific requirements of a protocol. This does not replace investigator judgment or Sponsor oversight, but it provides a clearer baseline for expectation-setting at the outset of a study.
In practical terms, it helps sponsors move from hoping a site will perform to having evidence-based confidence that it can.
Turning scale into intelligence
It’s fair to ask why this approach has not been widely implemented before. The answer is less about technology and more about data.
AI is only as effective as the information it is trained on. Most site models are too fragmented, too inconsistent, or too lightly instrumented to support meaningful performance analysis at scale. Velocity’s integrated network structure gives us access to deep, standardized, longitudinal data across sites, investigators, and therapeutic areas. This is when a site network scales for a purpose, turning operational insights into intelligence.
Because our sites operate within a shared framework with a common tech stack, we can compare performance on a like-for-like basis and identify patterns that would otherwise remain invisible. That capability allows us to deliver value not just through footprint, but through insight.
Raising expectations across the network
An important and often overlooked benefit of this approach is how it supports continuous improvement among sites and our personnel. By making performance data more visible and actionable, the tool helps identify where specific site capabilities can be strengthened. Sites that may not be optimal for one protocol gain clarity on how they can improve their alignment for future studies.
In this way, AI-driven feasibility becomes more than a selection mechanism. It becomes an addition to our quality framework that encourages transparency, accountability, and investment in operational excellence across the network.
What comes next
The first phase of this tool launches in February following extensive internal testing. To date, we have validated data sources, performance metrics, and analytical outputs by running real protocols through the system internally. While client-facing deployment is the next step, we have been deliberate in ensuring the foundations are sound before scaling use.
As with any AI system, its predictive value will increase over time. Each additional protocol assessed strengthens the model, creating a feedback loop that improves future feasibility decisions. While certain therapeutic areas, such as vaccines and metabolic disease, benefit earlier due to data depth, the longer-term impact will be felt across a broader range of indications.
A step toward a more predictable industry
AI-enabled feasibility will not eliminate the need for relationships or experience in clinical research. What it can do is support those relationships with stronger evidence and clearer expectations. In an environment where Sponsors are under pressure to reduce uncertainty and raise the bar on data quality, that combination matters.
For Velocity, this is part of a broader commitment to building a more predictable, data-led clinical research ecosystem. We encourage Sponsors and CROs to engage with us, understand how this approach works, and explore how data-driven feasibility can support their next study.
Because in today’s clinical research landscape, confidence is essential.
