Implementing AI in workflows: One step at a time

The OCT Europe 2026 Conference highlighted current challenges in therapeutic research and development (R&D) and discussed how novel technologies, especially AI, could overcome these inefficiencies.
AI-integrated workflows currently rely on a human-in-the-loop approach but aim to progress towards a human-on-the-loop strategy, emphasizing the importance of AI credibility.
Effective and ethical integration of AI into clinical trial workflows was a significant theme, with a focus on patient screening, data mining, molecular matching, and document intelligence as “real” applications of AI.
Small, incremental gains in less burdensome and repetitive processes are recommended for early AI implementation, cautioning against the “pilot trap” where disconnected pilot projects fail due to weak management and governance.
Training employees to effectively use AI-based tools is crucial but can be resource-intensive, particularly for smaller companies. This highlights the broader challenge of scaling AI deployment.
Transparency and consent are critical in maintaining patient trust, especially considering AI’s potential for algorithmic bias and the remaining low trust in pharmaceutical companies. Clear disclosure of AI use in clinical trials is emphasized.