SOC 2 + AIUC-1: what the joint AICPA report adds

A plain SOC 2 report tells you whether a service organization’s controls were designed and operating effectively. It was not written for AI agents. The joint SOC 2 plus AIUC-1 engagement, announced by the AICPA and AIUC in October 2026, adds a scoped certification of specific agents, including live behavioral testing. This page sets out what each half covers, who signs what, and what the combined report still does not tell you.

What was announced

The AICPA’s release is dated NEW YORK and SAN FRANCISCO (Oct. 19th, 2026). It says the AICPA and AIUC are rolling out a joint assurance certification that provides the first comprehensive, independent review of business risk related to AI systems and agents. The problem it targets is the vendor questionnaire: Trust is negotiated one security questionnaire at a time: a slow, costly exchange that rarely settles the question and hinders AI adoption.

On timing, the release says Several CPA firms are now delivering a joint SOC 2® plus AIUC-1 offering to clients. and that, Once initial engagements have concluded, a full launch of the joint certification is expected to take place in early 2027. AIUC’s own page names the firms: SOC 2 + AIUC-1 is available with BDO, Deloitte, EisnerAmper, EY, Grant Thornton, KPMG, PwC and Schellman.

Who does what

AIUC’s accreditation page lists some firms as provisional and says Provisional accreditation means the auditor is currently undergoing a witness audit. Check the firm’s status on that page before you rely on a report.

What a plain SOC 2 covers

The AICPA describes SOC 2 as an independent examination performed by a licensed CPA under the AICPA Attestation Standards that evaluates controls relevant to the security, availability, processing integrity, confidentiality or privacy of a system. The resulting report includes an opinion on the design and operating effectiveness of controls supporting the system.

The criteria come from the 2017 Trust Services Criteria, which the AICPA says presents control criteria established by the AICPA’s Assurance Services Executive Committee (ASEC) for use in attestation or consulting engagements. AIUC’s crosswalk summarizes the structure with security as the required baseline, and availability, processing integrity, confidentiality and privacy as optional extensions. AIUC’s FAQ says of SOC 2 that it is not AI-specific and therefore doesn’t cover risks unique to AI.

What the AIUC-1 half adds

• A narrower scope. AIUC says SOC 2 covers the organizational systems within its audit scope. AIUC-1 covers the specified agents or platforms and their security, safety and reliability risks. The joint engagement’s AIUC-1 scope is Scope defined in the Management Assertion - limited to AI agent systems, and a certificate is Typically limited to specific product(s), not the full organization’s products and practices.
• Live testing. The AICPA says What sets AIUC-1 apart is live testing of AI agent behavior, not just a review of policies and guardrails. AIUC’s Q4 2026 update requires every certified agent to be evaluated on a quarterly basis, with typical evaluations covering 1,000-3,000 simulations.
• Two deliverables. The joint offering lists a Full SOC 2 attestation report of selected SOC 2 TSCs and a Full AIUC-1 certification report of applicable AIUC-1 requirements including technical evaluation results of agent behaviour.
• One evidence request. AIUC says Both frameworks are scoped upfront and pursued in the same audit window and Overlapping requests are only submitted once.
• A shared control set. The Q4 2026 update says that, Across the 78 core controls in the 2026 Q4 standard, 40 are met by the deployer. AIUC adds that The responsibilities outlined in the guide are not legally binding.

Renewal and versions

The AICPA says both SOC 2 reports and AIUC-1 certifications require annual renewal, in addition to quarterly AIUC-1 technical retesting. AIUC’s FAQ says The AIUC-1 certificate is valid for 12 months. Evals are required at least every 3 months to keep the certificate valid. The standard itself moves: AIUC-1 is updated quarterly, with releases on January 15, April 15, July 15, and October 15 of each year. The October 2026 release says This quarter’s standard update led to 5 new requirements, 5 new controls and 3 retired requirements.

When you receive a report, note which quarterly version it was audited against and when the last quarterly eval ran.

What it deliberately leaves to SOC 2

AIUC says the standard avoids duplicating the work of non-AI frameworks like SOC 2, ISO 27001, or GDPR. Its SOC 2 crosswalk marks the gaps on its own side:

• AIUC-1 has no capacity or availability requirement, so there is no overlap
• AIUC-1 has no board-governance requirement, so there is no overlap
• system recovery/restoration (backups, failover, restoration procedures) is not required explicitly by AIUC-1

An AIUC-1 certificate on its own does not cover those topics. In the joint report, the SOC 2 half does, but only for the Trust Services Criteria the vendor selected.

What the certificate does not promise

AIUC states that an AIUC-1 certificate cannot guarantee AI system security, safety and reliability, and that certification represents a point-in-time assessment of controls, not a warranty of future performance or outcomes. AIUC also lists organizations Primarily using AI chatbots like Copilot, Claude or ChatGPT among those for which the joint offering is not relevant. It is aimed at companies that build or deploy their own agents.

TQA 9561: a separate AICPA track

On 11 September 2026 the AICPA published TQA Section 9561. The Q&As are intended to address the effect of a service organization’s use of AI on a SOC 1 examination performed in accordance with AT-C section 320 or a SOC 2 examination performed in accordance with AT-C section 205, Assertion-Based Examination Engagements. The AICPA’s 13 October 2026 release on its audit practice aid lists Guidance on the effect service organizations use of AI may have on SOC 1 and SOC 2 engagements.

TQA 9561 is guidance for auditors performing an ordinary SOC examination of a company that uses AI. It is not the AIUC-1 certification. The full TQA text requires an AICPA login, so we do not summarize its individual questions.

For safety-critical buyers (editorial judgment)

A joint report is useful evidence about an agent vendor’s security and reliability controls, backed by test results. It is not functional-safety evidence. No AICPA or AIUC page we cite claims it replaces DO-178C, IEC 61508 or ISO 26262 assurance, or that it shows conformity with the EU AI Act. If you need AI management-system evidence at organization level, ISO/IEC 42001 certification is a different scope.

Questions

Does the joint offering produce one report or two? The AICPA calls it a single CPA-led engagement. AIUC lists a full SOC 2 attestation report and a full AIUC-1 certification report as deliverables.

Who issues the AIUC-1 certificate in a joint engagement? AIUC. The CPA firm issues the report; AIUC says only it can issue the certificate.

Is the joint certification generally available? As of 25 October 2026, the AICPA says several CPA firms are delivering it, with a full launch expected in early 2027.

Is TQA 9561 the same thing? No. It is AICPA guidance on how a service organization’s use of AI affects an ordinary SOC 1 or SOC 2 examination.