Top Trends and Takeaways From the 2026 ACDM AI Symposium 

Last year, much of the conversation around AI in the life sciences focused on what might be possible. At the 2026 ACDM AI Symposium, held in London, the emphasis had clearly changed: Sponsors and vendors were talking about what they had already deployed, what was working, and what still stood in the way. Even in the fast-paced world of technology invention, seeing this type of innovation and adoption occurring so swiftly is unusual and incredibly exciting.

Even more exciting is that the accelerated adoption of AI is showing real results. Sponsors are deploying real use cases and seeing real efficiencies, while vendors are building at pace. That said, the atmosphere at the conference wasn’t one of unchecked optimism. Alongside every success story were practical questions about governance, workforce changes, and whether organizations have the right data foundations to support AI at scale.

Those conversations surfaced four trends that stood out to me throughout the symposium.

Takeaway 1: Innovations Are Leading to Efficiencies

While there were some moments of AI skepticism present during post-presentation Q&As, most presentations themselves centered around AI accomplishments. Sponsors presented live deployments, detailing how internal SOPs are being automated, medical coding is being streamlined, and clinical trial design is being accelerated. Roche built an internal chatbot for process questions. Deloitte is building AI platforms for trial design.

But while organizations are restructuring their roles, processes, and oversight to let AI do what it’s good at, discussions at the conference signaled that the bottleneck has shifted. AI can work, but can organizations adapt their technology and processes so that they are leveraging AI in ways that are both safe and effective?

Takeaway 2: The Role of the Data Manager Is Evolving

A popular AI narrative among skeptics is that AI is going to replace jobs. At the symposium, the story vendors and sponsors were telling indicated that data managers aren’t going anywhere, but their jobs are fundamentally changing. AI is now capable of handling laborious and routine work, which is freeing up capacity for data managers to focus on tasks that actually require judgment—tasks like reviewing what AI systems flag, identifying what matters, and making the call on what gets escalated. This was heartening to hear; however there’s tension underlying this AI reality as well. In one of the sessions I attended, someone asked: If data managers stop doing the routine work, won’t they lose some of the foundational skills they need to be able to do higher-level reviews?

My take: As organizations integrate AI, they must consider the type of training data managers will need to be effective humans in the loop.

Takeaway 3: The Tension Around Governance Is a Good Sign

The governance around AI is still being written. For me, a standout presentation was one from the MHRA signaling the reality of AI governance: There’s no rulebook yet, but one is coming. Their message was that anything produced by AI needed to be reviewed and heavily edited. They were incredibly direct in stating that if you use AI in your submissions, they’ll know, and they won’t accept the work. They want organizations to be deliberate about where AI sits in their processes and be able to explain the way it operates.

The tension that arose from the governance discussions was incredibly productive. It signaled that regulators have accepted the usefulness of AI and are thinking carefully about how it can be safely deployed. One of the main to-dos now is to figure out at what stage in the process the human in the loop belongs.

Takeaway 4: Data Readiness Is Essential

The uncomfortable truth that kept surfacing throughout the symposium: AI is only as good as the data and processes you feed it.

There was a great deal of enthusiasm and discussion around the possibilities found in agentic AI. But the uncomfortable truth was that any AI agent you deploy will only be as good as the data it can access. And since many organizations continue to operate siloed data systems, agents aren’t yet able to act to their full potential.

In one of the sessions I attended, a speaker discussed the data cleaning they had to do before launching their AI pilot. They described the effort as “the hard work…the groundwork…the unsexy and unglamorous stuff.” In other words, data cleaning is a laborious job.

But I think another truth that needs to be surfaced is that the work to make data accessible to AI agents doesn’t have to be hard. The bottleneck here is an organization’s data foundation. For an AI agent to create efficiencies in the clinical trials process, organizations need data that’s unified, not siloed across disconnected systems. Organizations need to rethink their current tech stack and focus on finding a platform that centralizes data collection and management, standardizes how data is captured and stored, and creates the governance and audit trails that make every decision traceable.

Carrying the Conversation Forward

Together, these trends suggest that the next phase of AI adoption will be defined less by what technology can do and more by whether organizations have the data, governance, processes, and skills to use it responsibly.

So, what should sponsors and CROs be doing right now? Given how quickly AI innovations are emerging, organizations that want to stay competitive will want to start building a data foundation that will support both current and future AI projects. Keep an eye out for regulatory guidance and governance but, in the meantime, establish internal governance and AI-use guidelines. Organizations that start this work now will be well positioned to deploy AI safely and effectively. Those that don’t may very well get left behind.

Interested in building a stronger foundation for AI in clinical trials? Explore the elluminate® Clinical Data Intelligence Platform and see how unified, governed clinical data can help you accelerate AI initiatives with confidence.

Shaun Chard

Business Development Director

Connect with me on LinkedIn.