Biopharma leaders are exploring LLM foundation models to support in-licensing and portfolio decisions, yet the true performance of these tools in high-stakes settings remains unclear.
In this original research article, Scitaris benchmarks Copilot, Claude, and Gemini against 40 expert in-licensing assessments and quantifies how often model recommendations diverge from ground truth.
Senior consultant Wignand Mühlhäuser, PhD, and Scitaris co-founder Jonathan Vonnemann, PhD, show where current LLMs misjudge opportunities even after prompt engineering. They also explain how Scitaris combines curated knowledge and vetted LLM workflows to keep decision quality under tight control while still unlocking value from AI.
