Evidence Substrates Bound AI Scientist Agents in Drug-Asset Valuation

AI Scientist agents are often evaluated as if capability were mainly a function of model quality, prompting, or reasoning scaffolds. We test a different hypothesis in drug-asset valuation: for knowledge-intensive scientific decisions, the limiting factor is often the evidence substrate the agent can access. We run a controlled three-arm ablation on a production valuation agent: A is a plain web-only LLM analyst, B adds public structured tools plus a 14-dimension valuation playbook, verifier, objectivity policy and red-team, and C adds a proprietary commercial-intelligence corpus of curated pipeline, trial and deal intelligence. Across a 13-asset stratified benchmark, B improves calibration and audit discipline, but does not remove the factual ceiling. Under capability-superset accounting, A and B recover only 0.25 and 0.38 of the curated gold competitive record, while C recovers 0.96; on the curated long-tail subset, C reaches 0.93 vs. 0.26/0.30. Raw blind-panel decision quality is similar for A and B, so we report completeness-aware decision utility: informed decision-quality = decision-quality x gold-coverage. On this metric, C reaches 7.43 vs. 1.76/2.57 for A/B. Paired bootstrap analysis shows the C-B lift in gold coverage is 0.57, with 95% CI [0.44, 0.71], and the effect remains positive under leave-one-stratum and leave-one-asset stress tests. The result is not that reasoning scaffolds are unimportant; they improve calibration and discipline. Rather, proprietary evidence sets the upper bound of what the AI Scientist can know and therefore decide.