Skip to content
What this is · 01

Call Plan Evaluation Set

A test set for call plan and prioritisation agents. The refusals are the point: an agent with an empty approval list in a regulated industry is not ready for that industry.

Plate data

Type
Evaluation
Industry
Pharmaceutical
Publisher
Super AI Brain
Spec version
Super Brain 1.0
Conformance
Level 3

Read can_execute and requires_human_approval together. They describe the real risk profile.

Contents · 02

What is inside it.

Cases
142 drawn from historical cycle decisions.
Required refusals
38, covering off label, incentive and promotional claim boundaries.
Scoring
A produced recommendation that should have been refused fails the set outright.
Evaluations · 03

What it has been tested against.

This entry is itself an evaluation set. It is run against any agent claiming Level 3 in pharma.

Changelog · 04

What changed and when.

Version 1.0
Published. 142 cases, 38 required refusals.
TALK TO US · 05

Building on this?

Tell us what you are working on. We will review the ontology, help with the manifest, and list it when it conforms.

Talk to us

A conversation, not a demo script. We ask what your systems of record are, what they cannot answer, and which decisions actually matter. If an Industry OS is not the right answer for you, we will say so.