2026 SCRS Global Site Solutions Summit: What's the Biggest Mistake Sites or Sponsors Make When Implementing Advanced Al?

Mike Wenger, chief innovation officer, CRIO, identifies two opposite failure patterns that derail sites and sponsors moving from basic AI tools to more advanced implementation.

The first, according to Wenger, is fear-driven avoidance, where sites believe they aren’t permitted to use AI at all out of concern it will violate sponsor expectations. Wenger pushes back directly on this, stating sponsors have never mandated that sites avoid AI; rather, they expect contracts and confidentiality to be honored while using it responsibly, and sponsors are using AI themselves. He encourages sites operating under this assumption to challenge that thinking, proactively engage sponsors, and arrive with a well-considered plan rather than waiting to be asked.

The opposite failure, in his view, is overcorrecting toward aggressive adoption, such as cutting staff or treating an AI agent as a literal replacement for a coordinator. Wenger argues this introduces real risk because AI performs certain tasks very well but not consistently, functioning as something of a black box; a pilot can look flawless in a controlled demo yet fail unpredictably, “crashing the car,” once deployed broadly, without a clear explanation for why.

His underlying principle is that AI should be scoped narrowly, handling small, well-suited pieces of a workflow rather than entire processes. He distinguishes this from deterministic, traditional software, where identical inputs reliably produce identical outputs and carry little risk. AI is better reserved for tasks like interpreting ambiguous language in a document, while core, repeatable processes should remain in traditional, deterministic code where reliability matters most.