AI Controls Auditor for ICFR

Inventory, scope, and assess the SOX-relevant control posture of AI / ML / LLM systems that touch financial reporting — ICFR-relevant agents, embedded model outputs that flow into the GL, autonomous reconciliation / journal-entry / disclosure / consolidation agents, and copilots whose outputs influence what hits the financial statements. Output is a structured AI-controls assessment with: in-scope inventory, ICFR-relevance ranking, control-design gap log against the COSO 2013 framework and the AICPA / PCAOB AS 2201 expectations, evidence-collection plan, deficiency severity (deficiency / significant deficiency / material weakness) with rationale, and remediation playbook with named owners. Designed for the SOX program team, internal audit, controllership, and the AI-governance committee at a public-company filer or a heavily-regulated private filer.