For sponsors and CROs running clinical trials, agentic AI might be as scary as it is compelling. There’s no end to its potential to create efficiencies and improvements in i
That's changing quickly as two viable paths emerge for responsible AI in clinical trials: vendor-driven AI in a closed system, or user-led AI in an open, governance-enabled system.
In the vendor-driven approach, some clinical data platforms are building AI directly into their software, a prescriptive, easy-to-understand approach, but one that constrains your AI to the vendor’s system and development path. s use cases multiply across your organization and new AI models evolve at lightning pace, this constraint may ultimately limit how far your organization takes AI.
Alternatively, there's an open model in which a governance layer allows an organization's AI to safely interact with clinical data system, letting users and organizations dictate AI uses.
Walker Bradham, head of product and design for Zelta, explains both approaches, including an open approach Zelta has developed with a model context protocol connector that enforces enterprise compliance while enabling deep AI flexibility.