Evaluating in-house AI for target ID and lead optimization: practical insights

[Eleanor.L]

I'm a computational biologist at a mid-sized biotech firm, and our team is evaluating whether to invest in building an in-house AI platform for early-stage drug discovery, specifically for target identification and lead optimization. The potential seems enormous, but I'm skeptical about the real-world applicability beyond the published hype and concerned about the data quality and infrastructure required. For others working at this intersection, what has been your practical experience with AI in drug discovery? What are the most tangible successes and frustrating limitations you've encountered, and what criteria would you use to decide between partnering with a specialized AI vendor versus developing proprietary models?