.png)
This article was originally written for and published by Drug and Device World.
Artificial intelligence (AI) is reshaping industries across the globe, but in MedTech, its adoption is more cautious and complex than in other sectors. While high-tech firms can afford to iterate quickly, medical device companies face unique challenges: stringent regulatory oversight, long product lifecycles, and the reality that lives are on the line if technology fails.
Despite these constraints, momentum is building. AI is moving beyond pilot projects into real-world applications across diagnostics, manufacturing, quality, and even regulatory processes. Propel’s State of Product Innovation 2025: AI Adoption and Collaboration Trends highlights this shift, showing that 65% of companies already report success using AI in product operations, with more planning to scale up investment.
To better understand how MedTech companies can strike the right balance between opportunity and risk, Drug and Device World spoke with Chuck Serrin, Vice President of MedTech & Life Sciences Industry Marketing at Propel. Serrin offers a grounded view of how AI is being adopted, where it is already driving impact, and what must change for MedTech to fully embrace this technology.
Why MedTech Moves Slower Than High Tech
Compared to high tech, where AI adoption has surged, MedTech still lags, sometimes by as much as 50%. For Serrin, this was no surprise:
“The risk is so high for a medical device company. Lives are at stake, and the cost of a recall or even a single patient incident can be devastating, especially for smaller firms. That makes companies wary of adopting technologies that regulators or clinicians might not fully understand.”
Propel’s research supports this divide: while 51% of high-tech firms plan to increase AI investment, only about a third of industrial and a quarter of MedTech firms are accelerating at the same pace. The difference is not about ambition, Serrin emphasizes, but about context. High tech can afford to “fail fast.” MedTech cannot.
The long product lifecycles typical of MedTech, often spanning years rather than months, add to the friction. Regulatory scrutiny at every stage means there is little tolerance for “black box” algorithms whose decision-making cannot be explained.
Early Wins
Despite these challenges, certain AI applications have already gained traction in MedTech. Radiology is a standout example, with companies using AI to analyze CT scans or detect anomalies. In these cases, AI provides a second set of eyes before specialists confirm results, creating both efficiency and reassurance.
Serrin sees similar momentum in digital therapeutics, particularly in behavioral health.
“We’re seeing AI help scan brain images to assess depression levels or identify which areas of the brain are affected. An expert then validates the output, ensuring safety. In some cases, the AI-powered software itself is the medical device. In others, software guides a device or drug that delivers treatment. This convergence of software, device, and therapy is accelerating because of AI.”
Such examples underscore how AI is opening doors to incremental adoption, rather than disruptive overhauls. Rather than reinventing the entire product lifecycle, companies are embedding AI into specific workflows where safety nets, like clinician oversight, can be built in.
Collaboration and Breaking Down Silos
Beyond clinical use cases, one of AI’s most powerful contributions lies in how it can foster collaboration across traditionally siloed MedTech teams.
“Quality is often off in one corner, regulatory in another, engineering doing its own thing. Each makes decisions in isolation, but when something goes wrong, the ripple effects are huge,” Serrin explains.
He describes scenarios where an engineering team might choose components not approved for certain uses, while quality departments remain unaware of recurring supplier issues affecting multiple product lines. Regulatory and marketing teams, often brought in late, then scramble to catch up.
This lack of coordination is reflected in Propel’s data, with the report revealing that 55% of companies cite lack of an integrated system as a major barrier to collaboration. Serrin argues that unified platforms, along with AI’s ability to surface connections across them, can close these gaps.
“If the regulatory team sees requirements earlier, they can shape clinical claims sooner. If marketing [team] is looped in, they can prepare collateral and labeling in parallel. Quality teams can catch systemic supplier problems before they hit multiple product lines. AI makes these connections visible, but only if the organization enables data to flow across teams.”
Listening Beyond Complaints
AI is also expanding how MedTech companies capture and act on customer feedback. Traditionally, adverse event reports and formal complaints have been the main inputs into product improvement. But Serrin points out that patients and clinicians share feedback in many informal channels.
“Someone might post on social media that a device didn’t work as expected. It’s not a formal complaint, but it’s still valuable. AI can scan these unstructured data sources to highlight emerging issues or unmet needs that companies might otherwise miss.”
He sees equal value in monitoring competitor complaints, which are often publicly available: “If a competitor’s device is failing in a certain way, and your device is predicated or built on similar principles, that’s a predictive signal. AI can surface it before it becomes your problem.”
This proactive, data-driven approach could shift MedTech from reactive compliance to continuous product innovation. It allows companies not only to protect brand loyalty but also to feed new requirements directly into next-generation designs.
Data Integrity and Trust in AI
At the foundation of every AI application lies a deceptively simple question: Is the data accurate and trustworthy? For Serrin, this is where MedTech companies cannot afford shortcuts.
“When product specifications are wrong, when suppliers build to the wrong revision, when data is spread across disconnected systems—that’s when recalls can happen. That’s when regulators lose trust.”
The cost of fixing errors late in the product lifecycle, during clinical trials or post-launch, is exponentially higher than getting it right at the earlier design stage.
Propel’s findings echo this: companies that invest in integrated product platforms report faster launches and fewer errors, with AI-driven tools improving both data quality and accessibility. AI helps, Serrin notes, but only if guardrails are in place:
“AI can pull vast amounts of data, but unless it’s curated and secured, it can lead to misuse. You don’t want sensitive product data leaking externally, or marketing teams making unverified claims. Role-based access and built-in safeguards are non-negotiable.”
Lessons From High Tech
While MedTech’s adoption curve is different, Serrin notes that there are important lessons to borrow from high-tech and industrial sectors.
One is predictive maintenance. Just as industrial IoT systems monitor equipment for early signs of failure, connected medical devices are now feeding real-time patient data to the cloud, where physicians can intervene sooner or advise more effectively.
Another is personalization. Serrin points to the growing use of AI in tailoring devices and therapies to individuals: “In the past, devices were commonly designed for a broad intended use, a ‘one size fits most’ approach. Now, with AI and 3D printing, you can design implants, dental crowns, or even cartilage tailored to a patient’s unique anatomy. Or AI-powered software for behavioral therapeutics that target optimal locations on the brain for stimulation. What started in high tech is now reality in MedTech.”
Such personalization not only improves patient outcomes but also aligns with broader healthcare trends toward precision medicine.
Building Momentum Through Incremental Wins
Given the risks and regulatory challenges, Serrin advocates for an incremental approach to AI adoption. Rather than aiming for moonshots, companies can seek small but meaningful wins that build trust and cultural acceptance.
He offers examples already in play:
- AI agents that automatically surface issues and complaint data from across platforms, helping quality teams act faster.
- Tools that generate training quizzes from new SOPs, reducing compliance burdens and saving time.
- Early deployment of AI to connect supplier, quality, and product data, spotting systemic risks.
“These may seem small, but they deliver quick results and build user confidence. Once teams see AI saving them time and improving accuracy, it’s easier to expand into bigger, more advanced applications.”
Propel’s report echoes this pragmatic strategy: companies that begin with smaller wins see faster cultural acceptance and more willingness to scale.
Looking ahead, Serrin is cautiously optimistic. He foresees regulatory clarity improving around AI/ML-based medical devices, especially as agencies gain more experience evaluating adaptive algorithms. He also expects integrated product platforms, combining design, regulatory, quality, and commercial data, to become the norm.
But perhaps most importantly, he stresses the need for explainable, transparent AI: “When outcomes affect patient health, black-box decisions aren’t acceptable. Clinicians, regulators, and patients must understand how an algorithm reached its conclusion. That’s the only way to build trust.”
Path Ahead
AI’s role in MedTech is no longer a matter of if, but how. The industry’s caution is justified, as lives are at stake, and regulation is exacting, but the early wins in diagnostics, quality management, and operations show what’s possible. Propel’s research confirms the trajectory: companies are moving from pilots to scale, provided they invest in collaboration and data integrity on secure and unified platforms.
For leaders like Chuck Serrin, the roadmap is clear: start small, build trust, break down silos, and aim for explainable AI. The future of AI in MedTech won’t be defined by hype, but by how effectively companies turn early wins into lasting impact.
.png)
