Presentation asked what clinical data management can safely delegate to AI — and what makes that delegation trustworthy
Trialt, a global clinicalresearch organization (CRO) specializing in biostatistics, statisticalprogramming, data management, and medical writing, announced that Chief R&DOfficer Yuri Kartashov, PhD, delivered a keynote presentation this week at COGCRO Summit Europe at Novotel Amsterdam City in the Netherlands. The event broughttogether senior leaders from across the European CRO sector to discuss technology,partnerships, resource optimization, site relationships and other prioritiesshaping clinical research.
Dr. Kartashov’s presentation, “Assistedto Automated to Autonomous: Evolution and Impact of AI in CDM,” distinguishedthree modes of working with AI: assisted, where AI drafts and people decide;automated, where a predefined workflow runs within agreed rules andcheckpoints; and bounded autonomous, where an agent chooses its next stepswithin limits people set and oversee.
Using a familiar example — a data querythat AI could draft, send to a site or close — he shows how each addedpermission changes the evidence a team must demand, checks recent AI claimsagainst documented outcomes in medical coding, data review, report generationand trial screening, and sets them beside documented failures and near misses.Findings from a pre-conference survey of CRO and sponsor data-managementleaders point the same way: drafting is widely trusted, sending and closingqueries are not, and supervised use rather than autonomy is what most expect bythe end of 2027.
“A query drafted in seconds can stilltake days to resolve,” said Dr. Kartashov. “The unit of value is an acceptedoutcome, including the time people spend checking, correcting and recovering.Permission should follow evidence: start with one repeated task, keep a personapproving each action, measure the work left for people, and expand only whenthe evidence supports it.”
The presentation also offered practicalguidance on setting up an agentic workflow for routine data management tasks,such as query drafting, data-transfer comparison and reconciliation: how togive an agent narrow, named tools with read-only access to study data, where toplace human checkpoints, how to bind each approval to the current data version,and how to test difficult cases before any live use. Dr. Kartashov outlined afirst AI-assisted script, a learning routine that stays with the work, and thehabits, skills and cost decisions that come with adopting AI. Attendees are invitedto leave with one recurring task to turn into a tested, reviewable workflow.
“Yuri has spent years turning technologyinto working solutions for clinical research, which is why his view of AI isgrounded in evidence rather than hype,” said Yaroslav Sud, VP of Biometrics,Trialt. “This talk gave data management leaders a practical way to decide whatto delegate now, what to hold back, and what evidence to ask for beforewidening AI’s responsibilities.”
COG CRO Summit Europe is the EU-focusedevent in the global COG CRO Summit Series and is designed specifically forsenior management and clinical operations professionals within CROs. Artificialintelligence for CROs is among the featured topics for the 2026 program.
For program information, visit https://www.thepbcgroup.com/cog-cro-summit-europe.To learn more about Trialt’s clinical data management expertise or to request acopy of the presentation, visithttps://trialt.com/clinical-data-management/.
About Trialt
Trialt is the trusted CRO partner forbiostatistics, statistical programming, data management, and medical writing.With expert teams across eight countries and deep engineering roots, we delivertailored solutions designed to move trials forward and produce measurableresults. Named “Company of the Year – Pharmaceutical Industry” in theInternational Business Awards, sponsors turn to Trialt to scale complexprograms, recover at-risk studies, and meet critical timelines with confidence.Learn more at www.trialt.com.
Media Contact
Don F. McLean, McLean Media
+1-734-716-4182
###