MedTech companies risk undermining artificial intelligence projects if they deploy the technology before establishing reliable data, standardised processes and clear governance, according to discussions at the Veeva MedTech Summit in Amsterdam.
Across presentations, panel sessions and interviews at the event, speakers returned to a common theme: the limiting factor is not how quickly an organisation can adopt AI, but whether the groundwork exists to use it effectively. That groundwork, as described at the summit, combines dependable data, consistent processes, suitable oversight and the people accountable for it.
Medtronic spent four years on readiness
Becki Brasher of Medtronic described a four-year effort to become ready to implement AI. The company did not begin with algorithms or tools. It first harmonised processes, improved data consistency, consolidated systems and established governance structures.
Brasher told the summit that Medtronic did not start with AI but had to earn it, framing the sequence as a precondition rather than a delay. The reasoning is that introducing AI into fragmented or inconsistent environments can create more problems than it solves: without reliable information and standardised processes, outputs are difficult to trust, validate or scale.
Medtronic's experience also pointed to workforce readiness. Technology alone does not deliver transformation, and employees need the knowledge, confidence and oversight to incorporate new tools into their everyday responsibilities.
Governance built in from the outset
In the closing keynote, Erik Orzechowski and Benjamin Rochette argued that as adoption accelerates, companies must move quickly without compromising the quality, safety and regulatory standards expected of the sector. Their position was that governance should be embedded in technology development from the start rather than treated as a final compliance exercise.
That requires closer collaboration between IT, quality and regulatory functions. IT teams understand the technology and its implementation, while quality and regulatory professionals assess risk and determine the evidence needed to demonstrate compliance. The speakers also advocated proportionate governance: an AI application supporting relatively low-risk administrative work does not require the same level of control as one influencing patient safety or regulatory decisions. Clear responsibilities, early risk assessments and appropriate controls allow a more flexible approach that supports innovation while keeping obligations in view.
Human judgement and the trust gap
Speakers also stressed the importance of retaining human expertise. Brasher made clear her view that AI should assist human reviewers rather than replace them, and said human judgement would remain in place until organisations build confidence that what agents deliver is accurate.
AI can analyse large volumes of information, generate documentation and provide recommendations, but its outputs still require scrutiny. In a highly regulated environment, a convincing answer is not necessarily a correct one. During the closing keynote, relying on AI without understanding its underlying subject matter was compared with using a calculator without knowing the mathematics behind it.
The concern extends to the next generation of regulatory and quality professionals. If employees become accustomed to relying on AI before developing the necessary underlying knowledge, organisations could find themselves with fewer people capable of challenging inaccurate or inappropriate recommendations. The consensus was that AI should complement human expertise, with accountable individuals retaining responsibility for important decisions.
Pilots, value and pace
With an increasing number of potential applications available, organisations were urged to resist pursuing every opportunity simultaneously and instead identify areas where technology addresses existing operational challenges, with clear measures of success. Value indicators cited in the closing keynote included time savings, reduced process cycle times, less rework, improved quality and employee adoption.
Controlled pilot projects were presented as a way to understand where AI is most effective, retain appropriate human oversight and build confidence before expanding successful applications. Because both the technology and regulatory expectations will continue to evolve, governance structures need to adapt alongside new capabilities.
Orzechowski described the overall approach as slowing down in order to speed up later. Success, the summit suggested, will depend less on the number of AI initiatives an organisation launches than on its ability to turn promising applications into reliable, scalable business capabilities, supported by collaboration among manufacturers, technology providers, regulators and other industry stakeholders.