Build the decision case before the committee deck

The most dangerous sentence in a committee paper is often the cleanest one: “Management recommends advancing the program.” It conceals the work that matters. Which version of the asset was reviewed? What patient population was assumed? Which liabilities were unresolved? What would have changed the recommendation?

A document is a release, not a starting point

Teams often begin by asking for slides. That reverses the order of operations. A committee paper should be released from a completed decision case, much as a financial statement is released from controlled books. The case begins with a decision, an accountable owner and a date. It names the alternatives, the evidence cut, the approved methods and the thresholds that separate advance from defer or stop.

The logic is familiar in drug development. FDA guidance describes the Target Product Profile as a statement of the labeling concepts under development and links those claims to the studies intended to support them (FDA). An internal capital decision deserves the same discipline. The desired action must be tied to evidence capable of supporting it.

In ARiDA, a Decision Case retains the question, asset revision, assumptions, analytical outputs, dissent, approval state and reopening triggers. Asset 360 provides the current scientific and business view. The released document summarizes that case; it does not replace it.

Compute before narrating

Suppose a licensing committee is considering an oncology asset. The work should not begin with a market story. It should resolve the target and molecule, assemble genetic and translational support, inspect on-target and off-target safety, establish clinical and regulatory precedent, define the commercial scenario, then calculate the consequences.

The valuation can include a reference-class outside view, correlated Monte Carlo rNPV, the probability of non-positive value, real-option value and value of information. If the result is bimodal, the mean receives a warning because it may describe no plausible outcome. If portfolio assets share scientific exposure, the covariance should reflect it. If a safety veto fires, the committee should see the veto before it sees an attractive aggregate score.

The result is a paper whose claims can be challenged precisely. A reviewer can dispute the source, the interpretation, the assumption, the method or the threshold. That is productive disagreement. Arguing over a polished paragraph with no visible calculation is not.

Reopening is part of the design

A decision is valid for an evidence state, not for all time. The case should specify what reopens it: a clinical hold, a competitor readout, a chemistry liability, a price change or a new experiment that resolves a key uncertainty.

This is also the practical meaning of governance. NIST’s AI Risk Management Framework calls for documented roles, evaluation, monitoring and lifecycle controls (NIST). In a biopharma decision system, those controls should be visible in the work product itself.

The committee paper is better because it is smaller than the case beneath it. It can be concise without being shallow.