By Sweta Pittala, VP, Technology Portfolio Management
At a Glance
Responding to Requests for Proposal (RFPs) is one of the most labor-intensive processes in clinical trial operations. Velocity’s Feasibility Agent automates and augments that process, drawing on years of data on site performance from 70+ global sites, PI experience, and local recruitment trends to produce ranked, evidence-based site recommendations in hours rather than weeks. Sponsors and CROs already receiving proposals from Velocity will have noticed the difference — and this is just the start.
Feasibility is one of the most consequential stages of a clinical trial. It’s also one of the most labor-intensive. For every new RFP, central and site teams have to interpret protocol requirements, assess sites against a complex matrix of criteria, and compile detailed responses to sponsor and CRO questionnaires, often under significant time pressure.
Velocity’s technology team is focused on making sure the network runs as effectively as possible, freeing colleagues to concentrate on delivering trials and serving patients rather than navigating administrative processes. Feasibility, with its high administration burden and direct commercial impact, was a natural fit for our agentic AI program. If you missed last week’s piece on how we’re deploying agentic AI across our operations, it’s worth reading alongside this one.
Introducing our Feasibility Agent
Feasibility has always demanded a lot from central and site teams; a detailed knowledge of the network, careful reading of protocol requirements, and a thorough understanding of what each site has delivered in the past. Our Feasibility Agent, rolled out last month, automates and augments that process.
The Agent reviews protocols and sponsor questionnaires, identifying and comparing study requirements to historic data on site performance, local recruitment trends, PI experience, and trial outcomes across our global network to produce ranked site recommendations. Each recommendation includes a probability score that indicates how well a specific site matches the trial requirements, allowing our Business Development teams to quickly identify the best sites for any given protocol.
What’s Behind a Recommendation?
Each recommendation draws on a huge corpus of data, assessing every site in the network simultaneously. Patient availability by condition is evaluated alongside local disease prevalence data from the CDC and EU public health sources, giving a picture of the eligible population within the site’s catchment area.
PI experience in that therapeutic area and a site’s historic trial performance are factored in, alongside any prior association between that investigator and the trial sponsor or CRO.
Site infrastructure, protocol-specific certifications, historic enrollment rates, screen fail ratios, and retention data from similar studies also contribute to the suitability assessment, as does a projected enrollment forecast modeled from performance history and calibrated to the study’s needs. This enables our teams to be sure that the sites ranked are not only the most suitable on paper, but have proven themselves capable of delivering in practice.
Sponsor and CRO Outcomes
Ultimately, it’s not only our internal teams that benefit from our Feasibility Agent. Higher-confidence matches at the feasibility stage also have downstream benefits for our customers.
Sponsors and CROs receive submissions faster and a clearer picture of the reasoning behind every site recommendation; not just which sites Velocity is proposing, but why, and what the data suggests they’ll deliver. If they do opt for Velocity, our partners can expect consistent outcomes in everything from enrollment rates to startup times, since they’re modeled from the outset on years of data.
If you’ve received a proposal from us in the last few weeks, you may already have noted the addition of site scores based on their match to your protocol and more evidence-based feasibility submissions.
The Agent is developing quickly and we’ve just launched an auto-generated feasibility response feature, built from a knowledge base drawn from years of historic submissions. After that, a strategic narrative layer that gives BD teams an AI-generated summary of Velocity’s competitive position for each opportunity. We’re also continuing to expand the factors feeding site prioritization as live usage informs how the model develops.
This is one part of a broader program of work for our technology team — building tools that make the network more consistent, more reliable, and more useful to the people running trials within it. The Feasibility Agent is an early example of what that looks like in practice, but we’re moving quickly… so watch this space.
