Latest quarterly update
Q3 2026
Practical impact
If you run a customer-facing agent
What this means for you If your agent doesn't generate or execute code, the headline coding-agent additions are unlikely to be your biggest Q3 issue. The more important change is scope. AIUC-1 clarified which controls apply according to an agent's capabilities and deployment context. That means an unchanged product does not necessarily mean an unchanged audit scope. Pay particular attention if your agent handles customer data, connects to third-party services, generates externally visible content, or has access to tools and systems. Q3 revised controls in each of these areas.
What to do
Re-check your Statement of Applicability against what your agent can actually do today — including integrations, data access and externally facing capabilities. Don't assume last quarter's control mapping still applies simply because the product itself hasn't materially changed.
If you're currently preparing for certification
What this means for you Don't respond to Q3 by rebuilding your entire evidence package. First determine whether the changed requirements and controls actually fall inside your audit scope. AIUC-1's Statement of Applicability is the definitive list of requirements and controls against which the agent is audited, and applicability depends on the system's capabilities, architecture and deployment context. For teams with code-generating agents, Q3 deserves particular attention: new mandatory requirements address secrets leakage and secure patterns in generated code, alongside expanded execution safeguards.
What to do
Ask one question before collecting more evidence: “Which Q3 changes alter our Statement of Applicability?” Resolve that with your auditor first. Then update evidence only for the requirements and controls that actually affect your scope.
If you're preparing for the first time
What this means for you You have one advantage: you don't need to retrofit an old AIUC-1 implementation. Start from the current release and determine scope before building an evidence library. AIUC-1 separates core and supplemental controls, and their applicability depends on the particular AI system. For coding agents, include secrets handling, secure code generation and execution safeguards in that assessment from the beginning. For other agents, map the system's real capabilities, data access, external integrations and deployment context before deciding which controls require evidence.
What to do
Create a simple inventory first: AI system → capabilities → data → tools/integrations → deployment context → applicable requirements/controls → available evidence. Only then start filling evidence gaps.
Official update
AIUC reports that the July 15, 2026 release updates 8 requirements and 41 controls. The release focuses on expanding coding-agent requirements, technical guidance for auditors, and clarifications to existing requirements.