At the European Drug Safety & AI Conference 2026, DrugCard CEO Dmytro Horilyk spoke about a challenge many pharmacovigilance teams quietly face when discussing AI adoption.
The industry conversation often focuses on models, algorithms, and automation potential. Yet the biggest barrier rarely comes from technology.
As Dmytro explained during the session:
“AI projects don’t fail because of technology. Typically, they fail because we don’t know how to buy change.”
Pharmacovigilance teams operate in an environment where every decision carries responsibility for patient safety and regulatory compliance. Because of that, the hesitation around AI often comes from accountability concerns rather than technical limitations.
“We always ask ourselves: what if the model is wrong? What if an auditor asks a question and I don’t have the answer?”
These concerns explain why many companies remain cautious. At the same time, staying fully manual creates its own operational risks as safety data continues to grow across global markets.
The Hidden Risk of Doing Nothing
Pharmacovigilance professionals are trained to identify visible risks in safety signals and case processing. A less visible risk appears when companies delay technological change for too long.
During the talk, Dmytro highlighted a key paradox within the industry:
“When AI breaks, it’s an incident. But the systematic inability to process data at scale becomes your liability.”
The volume of safety information from literature, case reports, and global monitoring continues to increase every year. Systems that rely entirely on manual work eventually reach practical limits.
Strategy Before Technology
Another common mistake appears when companies treat AI as a simple procurement decision.
“Amateurs talk about the gear. Professionals talk about the strategy.”
Selecting a tool without a clear implementation plan rarely produces measurable results. Teams that succeed usually start with small experiments, iterate quickly, and gradually integrate automation into existing workflows.
From Theory to Implementation
DrugCard recently supported the pharmacovigilance service provider Asphalion in implementing AI-powered literature monitoring across multiple products and international markets.
The project helped automate article screening while keeping experts in the loop, resulting in more than 94% faster literature screening.
Watch the Full Talk
In the full presentation from the European Drug Safety & AI Conference 2026, Dmytro Horilyk shares practical insights on evaluating AI vendors, preparing teams for implementation, and navigating the internal decisions that determine whether AI projects succeed or stall.
Watch the full session below:
