Regulatory and quality leaders are eager to leverage AI to optimize critical industry-specific workflows, streamline decision-making, and improve operational outcomes. However, deciding where to start can feel complicated given the highly regulated nature of medtech and the rapid pace of AI innovation.
At the recent Veeva MedTech Summit, regulatory and quality leaders from Alcon, Merit Medical, and PricewaterhouseCoopers (PwC) offered guidance and shared examples to help others deliver effective AI business results.
Before diving into use cases, it’s important to establish governance and ground rules for an effective AI program. Bryan Hunt, global head of digital quality and regulatory affairs at Alcon, described his team’s approach.
Alcon developed a governance model that unified quality and regulatory opportunities for AI strategy, technology, and data governance. Hunt and his team established processes, standards, and SLAs to align business, IT, and compliance stakeholders.
The success of AI initiatives is fundamentally linked to the integrity and maturity of underlying processes and information. Hunt emphasized that having trusted and compliant data in core systems of record is critical. “If you don’t have good, accessible data and are automating bad processes, you’re not going to have a strong AI model.” There’s no way to avoid the hard work of re-engineering processes through automation in core systems of record that drive industry best practices. In fact, this automation and standardization is critical to AI.
To drive AI-enabled business outcomes, it is crucial to invest in organizational change management. Like other technology-based transformations, AI requires teams to change how they work, align disparate areas of the business, and develop new skills. However, because GenAI in particular replicates some activities that were formerly owned by human knowledge workers, the change feels even bigger.
Hunt emphasized the importance of training teams and supporting them during the change. “Most people think of AI as something that just works…” but getting it to work requires proper guidance. For example, prompt engineering, the practice of writing instructions to guide GenAI toward useful outcomes, requires focused training so teams can use new tools effectively.
Hunt encouraged businesses to develop a set of criteria and a prioritization matrix to evaluate the business impact of AI use cases, whether that be revenue, quality and compliance, or cost-effectiveness.
According to Hunt, “Last year, we learned about 500 AI use cases and couldn’t tackle them all. We prioritized 10 projects that will each deliver more than $2M in savings and then picked the right AI tools for the job.”
Filtering projects by value and feasibility helps foster success for both financial and customer outcomes. A structured use case selection process also ensures adequate support–and that teams proposing AI projects are held accountable for the results.
Cory Marsh, head of regulatory affairs at Merit Medical, observed that overly ambitious AI projects–like trying to build a custom LLM in-house–are high risk and failure-prone. Instead, Merit evolved its approach by investing in projects that target one to two processes to show value, gain knowledge, and build momentum.
Given the rapid pace of technology change, smaller projects also avoid over-indexing on a particular technology and allow flexibility as new capabilities become commercialized. Merit prioritized five key workflows in their “crawl, walk, run” approach:
Marsh emphasized that smaller, more achievable projects show value in months rather than years and build on one another iteratively. This approach satisfies executive leadership’s need to “do something” without falling prey to long lead-time, high-risk projects and allow business and IT organizations to adopt an agile approach to AI adoption.
Lastly, Marsh highlighted that strong use case criteria and prioritization is paramount to properly selecting these fast-track projects, while the scope and speed of the projects helps with change management and organizational learning.
In addition to driving regulatory efficiencies, AI models can also reduce administrative burden when it comes to quality processes. Sam Venugopal, global quality lead at PwC, described how his organization is helping multinational companies streamline investigative conclusions with AI.
According to Venugopal, medtech companies spend on average millions each year on product complaints and non-conformance processing events. To address the manual and labor-intensive nature of these tasks, PwC worked with Veeva to develop AI solutions to automate the identification and classification of non-conformance events. These tools can also evaluate and confirm root causes for non-conformance and predict future quality risks to help teams make faster decisions, while keeping a human-in-the-loop to ensure robust compliance.
“These are extremely complex processes; AI analyzes large data sets to detect patterns, find anomalies, and quickly identify quality issues compared to manual investigations.” An AI-driven root cause analysis tool, for example, can reduce root identification from ~30 days to minutes, according to Venugopal.
In summary, Alcon, Merit, and PwC agreed that medtechs can maximize the value and effectiveness of real-world AI transformations by:
For more details about how to get started, watch “Alcon & Merit Medical: Evaluating AI in Regulatory Affairs” and “Alcon & PwC: Innovating Quality Management with Advanced Technologies and AI” on Veeva Connect.