Risk and governance professionals are being asked to own AI risk without a credential that tests it, and ISACA's two newest advanced credentials, AAIR for AI risk and AAIG for AI governance, are aimed squarely at that gap. Advanced in AI Risk (AAIR) is open for registration; Advanced in AI Governance (AAIG) is in a closed beta. Both are add-on credentials for people who already hold a qualifying certification or substantial experience. This guide sets out what ISACA publishes about each: eligibility, exam content and weightings, fees and maintenance, and what is not yet published for AAIG. It adds a preparation plan, decision guidance against ISACA's AAISM and AAIA, and where each fits on a GRC career path. It is for risk managers, CRISC and CISA holders, governance leads and managers building AI oversight capability. Status as checked on isaca.org on 27 October 2026: AAIR launched, AAIG in beta. Check current fees and requirements on ISACA's pages before registering.
Key takeaways
- AAIR is available now. Registration is continuous, with six months of eligibility to sit; the exam has 90 questions and is delivered at PSI test centres or by remote proctoring.
- AAIR is weighted to running an AI risk programme. Domain 3, AI Risk Program Management, is 42%; Domain 1, AI Risk Governance and Framework Integration, 37%; Domain 2, AI Life Cycle Risk Management, 21%.
- AAIR needs a qualifying designation. ISACA lists CISA, CISM, CRISC, CGEIT and CDPSE plus a long list of external risk, audit and accounting credentials, including CISSP, CGRC, CIA and PMI-RMP.
- AAIG is in beta. The beta application period has closed; eligibility is an active CGEIT or six years of enterprise governance experience including at least one in AI. Format and fees are not yet published.
- Maintenance is light but strict. AAIR needs 10 CPE hours a year and 30 over three years, related to AAIR, and an annual fee; failing these revokes the designation.
Who the certification is for
| AAIR: Advanced in AI Risk | AAIG: Advanced in AI Governance | |
|---|---|---|
| ISACA's stated audience | Experienced IT risk professionals who work across functions to respond to AI risk | Experienced governance professionals tasked with leading AI adoption at enterprise level |
| Status (27 October 2026) | Launched; registration and exam preparation materials available | Beta in progress; beta applications closed; mailing list for launch updates |
| Typical holder | Risk manager, second-line risk function, CRISC holder, IT risk consultant | Head of governance, CGEIT holder, AI governance board member or programme lead |
| Practice areas named on the landing page | AI Enterprise Governance; AI Life Cycle Risk Management; AI Risk Program Management | AI Enterprise Governance; Data and AI Architecture Principles; AI Enterprise Integration and Performance Monitoring |
How they relate to AAISM and AAIA
ISACA's AI family now has four advanced credentials, each anchored to a foundation discipline. AAISM (security management) and AAIA (audit) are covered in the ISACA AAISM and AAIA AI Certifications Guide; this page does not repeat them.
| If your job is | Choose |
|---|---|
| Identifying, assessing and treating AI risk across the business | AAIR |
| Setting enterprise AI direction, architecture principles and performance oversight | AAIG, once launched |
| Running the security programme for AI systems | AAISM |
| Auditing AI systems and controls | AAIA |
For all certification options, see the Cybersecurity Certifications Guide and the AI Security Certifications Compared Guide.
Exam format and domains
AAIR format
| Item | AAIR (ISACA, checked 27 October 2026) |
|---|---|
| Questions | 90 |
| Duration | Not stated on the pages checked; see ISACA's Exam Candidate Guide |
| Delivery | Computer-based at PSI testing centres or remotely proctored; residents of India, Mainland China and Hong Kong must test at centres |
| Registration | Continuous; six months of eligibility from registration |
AAIR exam content outline
| Domain | Weight | Sub-areas (ISACA) |
|---|---|---|
| 1. AI Risk Governance and Framework Integration | 37% | 1A AI models, frameworks, strategies and use cases; 1B organisational processes and alignment; 1C ownership, oversight and accountability; 1D policies, procedures and training; 1E regulatory compliance and legal considerations; 1F trustworthiness, ethics and societal implications |
| 2. AI Life Cycle Risk Management | 21% | 2A design, development or procurement, and documentation; 2B model training, testing and validation; 2C implementation, maintenance and decommissioning; 2D data and asset management |
| 3. AI Risk Program Management | 42% | 3A risk scenario identification and assessment; 3B risk treatment; 3C controls management; 3D risk metrics, monitoring and reporting; 3E AI supply chain and third-party risk; 3F AI incident response, BIA, business continuity and disaster recovery |
The landing page's practice area names differ slightly from the content outline's domain names; study against the outline. Domain 3 is where risk professionals already hold most of the method (scenarios, treatment, KRIs, third-party risk), so the AI-specific content to learn is mostly in domains 1 and 2: model life cycle, validation, data and AI regulation.
AAIG format
ISACA has not published question count, duration, weightings or fees for AAIG on the page checked. Review materials (an online course, a 12-month question database and a review manual) are listed as coming soon. Check the AAIG page for launch details.
Where identity fits: AAIR sub-areas 2D (data and asset management), 3C (controls) and 3E (supply chain) all reach AI agents, model service accounts and API keys to third-party models. An AI risk register that ignores which identities an AI system runs as, and what they can reach, misses one of the most direct paths to harm. See the Agent Identity and Access Guide.
Prerequisites and experience
| Credential | Eligibility (ISACA) |
|---|---|
| AAIR | One qualifying designation. ISACA: CISA, CISM, CRISC, CGEIT, CDPSE. Global: CRMP, CRMP-FED, CRMA, CERP, CRCM, CGRC, CISSP, CIA, ANAN CNA, CGMA, Canadian CPA, ACCA, FCCA, US CPA, Japanese CPA, ACA, FCA, CA ANZ, FCA ANZ, CPA HKICPA, FCPA HKICPA, PMI-RMP. The page also references proven experience in IT risk or advisory roles. |
| AAIG | An active CGEIT, or six years of enterprise governance experience including at least one in AI |
For AAIR, after passing you must pay a US$50 application processing fee, adhere to ISACA's Code of Professional Ethics and CPE policy, and apply within five years of passing. Confirm the experience route in ISACA's candidate materials before relying on it, since the landing page states the designation list first.
Decision guidance: a CRISC holder already meets AAIR's eligibility and its domain 3 methods, so AAIR is the natural next step. A CGEIT holder should wait for AAIG if their role is enterprise direction; if their role is mainly risk, AAIR is available now. Someone with none of the listed credentials should first take the foundation one that fits their job; see the CRISC Certification Guide and the CGEIT Certification Guide.
Preparation plan
- Confirm eligibility (week 0): check your qualifying designation is active and listed.
- Register and plan the window: eligibility lasts six months from registration, so register when you can commit to a date.
- Baseline against the outline (week 1): rate yourself on each of the 16 sub-areas. Most risk professionals will be strong on 3A to 3D and weaker on 1A, 2B and 2C.
- Learn the AI-specific material (weeks 2 to 6): AI model types and use cases, the AI life cycle, training and validation risks, data provenance, and frameworks such as the NIST AI RMF and ISO/IEC 42001. See the NIST AI RMF Guide and the ISO/IEC 42001 Guide.
- Apply it to real cases (weeks 4 to 7): write three AI risk scenarios for your organisation (for example a customer chatbot, a code assistant and an AI agent with system access), with treatments, controls, KRIs and supplier risks.
- Cover regulation (alongside): sub-area 1E covers regulatory compliance; build a short map of AI rules that apply to you. See the EU AI Act Compliance Guide.
- Practise (weeks 8 to 9): use ISACA's preparation materials where available and time yourself against the duration in the candidate guide.
Small vs large organisations: in a small organisation one risk lead may own all AI risk, and AAIR gives a structured method quickly. In a large organisation, AAIR suits the second-line risk function, while AAIG (once launched) suits the governance office and AAIA the internal audit team, giving three lines a shared AI vocabulary.
Measuring readiness and value
- Coverage: all 16 AAIR sub-areas rated amber or green.
- Applied output: three documented AI risk scenarios with treatment plans.
- Practice scores comfortably above ISACA's published pass mark for the exam.
- After passing: share of AI systems in your inventory with an assessed risk, named owner and KRIs.
Cost, renewal and continuing education
| Item | AAIR (ISACA, USD, checked 27 October 2026) | AAIG |
|---|---|---|
| Exam registration | US$459 members; US$599 non-members | Not yet published |
| Application processing fee | US$50 after passing | Not yet published |
| CPE | At least 10 hours a year and 30 over a three-year period, related to AAIR | Not yet published |
| Annual maintenance fee | US$45 members, US$85 non-members, due by 1 January; US$25 and US$50 for a third or later ISACA certification | Not yet published |
| Double counting | CPE can count towards more than one ISACA certification where the activity applies to each | |
| Non-compliance | Revocation of the designation |
Fees change; check current fees on ISACA's pages before registering.
Common failure patterns
- Treating AAIR as CRISC with AI added: domains 1 and 2 test AI life cycle, validation and regulation that CRISC does not.
- Ignoring the six-month window: registering too early can waste the fee.
- Studying the landing page, not the outline: the outline's domains and weightings are what the exam tests.
- Forgetting maintenance: missing CPE hours or the 1 January fee leads to revocation.
- Waiting for AAIG when AAIR fits: if your role is risk, AAIR is the closer match and is available now.
Where it fits on the career path
| Foundation | Advanced AI credential | Career direction |
|---|---|---|
| CRISC, CGRC, CIA or another listed risk or audit credential | AAIR | AI risk lead, second-line risk, risk consulting |
| CGEIT or long governance experience | AAIG | AI governance lead, governance office |
| CISM or CISSP | AAISM (or AAIR for risk roles) | AI security management |
| CISA | AAIA | AI audit |
- See the GRC and Risk Certifications Compared Guide and the ISACA AAISM and AAIA AI Certifications Guide.
ISACA does not publish pass rates or salary figures on the pages checked.
Practitioner checklist
- Confirm your qualifying designation for AAIR, or CGEIT or governance experience for AAIG.
- Register for AAIR only when you can sit within six months.
- Study against the AAIR content outline, weighting time towards domains 3 and 1.
- Learn the AI life cycle, validation and data risks in domain 2.
- Write real AI risk scenarios, including AI agents and their identities.
- Map AI regulation that applies to your organisation.
- After passing, apply within five years, then log 10 CPE hours a year and pay the fee by 1 January.
- Watch the AAIG page for launch details if governance is your role.
Standards and references
- ISACA: Advanced in AI Risk (AAIR)
- ISACA: AAIR Exam Content Outline
- ISACA: Maintain AAIR Certification
- ISACA: Advanced in AI Governance (AAIG)
Related NHI Mgmt Group resources: ISACA AAISM and AAIA AI Certifications Guide · Cybersecurity Certifications Guide · CRISC Certification Guide · GRC and Risk Certifications Compared Guide · NIST AI RMF Guide