GAICC AI Conference & Awards 2026 "Governing the Future – Building Responsible, Safe and Human-centric AI"

Buy tickets Now →

Log in / Sign up

Enroll Now

GAICC AAIGP — Agentic AI Governance Professional Certification

Conventional AI governance asks whether a model’s output is accurate, fair and lawful. Agentic AI governance asks a harder question: what is this system permitted to do, on whose authority, within what limits, under whose supervision — and how do we stop it?

AAIGP is the advanced practitioner credential for the professionals who have to answer that question in writing, and stand behind the answer.

Last Updated: 2026
4 Days
English
Certification Trusted by Professionals at:

AI Certification Council

4.8 / 5.0 Rating
27,143+

Students

Our training programs are CPD Accredited by CPDSO

GAICC AAIGP — Agentic AI Governance Professional Certification

  • Programme : 4 instructor-led days · 32 contact hours
  • Examination : 100 scenario-based items · 150 minutes
  • Pass Mark : 70% · scaled 100–500, 350 to pass
  • Standard : Built to ISO/IEC 17024
  • Validity : 3 years · 60 CPD hours per cycle
  • Delivery : AI-proctored online, worldwide

Exam Content Outline (ECO)

Download the GAICC-AAIGP Programme Course Outline for full details about domains, modules, eligibility, and the application process.
Download
Candidate Handbook

What Is the GAICC AAIGP Certification?

The GAICC Agentic AI Governance Professional (AAIGP) is an advanced practitioner credential for professionals who authorise, oversee, assure and shut down AI systems that take action in the real world.

It is the first GAICC credential written for hybrid human and autonomous-agent teams rather than for AI systems that only produce output. Where earlier generations of AI governance were built around a model returning a recommendation, AAIGP is built around a system that books the flight, moves the money, files the ticket, publishes the document and calls the next agent — often without a human in the loop for any single step.

A credential for systems that act, not just answer

Every one of the 32 contact hours is spent on the four problems that autonomy creates and that no prior governance practice fully solves:
  • Authority — what an agent may do, on whose behalf, and how that grant is bounded, evidenced and withdrawn
  • Control — how oversight stays meaningful at machine speed and machine scale
  • Assurance — how you prove an agent behaves acceptably before you authorise it, and continuously after
  • Evidence — how you reconstruct an autonomous decision months later to a standard that satisfies an auditor, a regulator or a court

Built on the GAICC AI Governance Framework

AAIGP extends the GAICC AI Governance Framework, Maturity Model and Implementation Roadmap into agentic operation. It assumes the fundamentals taught in the GAICC Certified Professional in AI Governance (CPAIG) and does not re-teach them, which is why every hour goes to autonomy-specific competence.

The credential is issued under ISO/IEC 17024, the international standard for bodies certifying persons. Certification decisions are taken by the GAICC Certification Decision Authority, independently of the platform that scored your examination and independently of whoever delivered your training.

Frameworks and regulations applied throughout the programme: EU AI Act (Article 14 human oversight and GPAI obligations) · ISO/IEC 42001 · ISO/IEC 42005 · ISO/IEC 23894 · NIST AI RMF · GDPR and equivalent global privacy regimes (PIPL, PDPA, LGPD) · OECD AI Principles · the GAICC AI Governance Framework

Enroll Now

GAICC-AAIGP Certification

Issued By Global AI Certification Council

Prepare for your AAIGP Certification with this course

Why Agentic AI Governance Is Now a Board-Level Problem

Autonomy has outrun the governance model

Most organisations approved their AI governance framework when AI meant a model that produced text. That framework asks whether output is accurate, fair and explainable. It has no answer for a system that holds credentials, invokes tools, spends money, writes to production systems, calls other agents and completes a transaction end to end before a human sees it.

The result is a widening gap between authority granted and assurance held. AAIGP exists to close it.

Regulation is silent where agents are loudest

Current AI regulation was drafted for systems that inform decisions, not systems that execute them. Practitioners are routinely asked to make defensible calls in areas where the law has not yet spoken — liability for an autonomous action, notification duties when the actor is an agent, the point at which agentic customisation converts a deployer into a provider.

AAIGP trains you to reason from first principles where regulation is silent, and to say so honestly rather than over-claim compliance. That is why no exam question can be answered from vendor terminology, and no question tests definition recall alone.

AAIGP vs CPAIG vs general AI governance credentials

Dimension GAICC AAIGP GAICC CPAIG General AI Governance Credentials
Governing question What may this system do, and how is that authority bounded and evidenced? Is this AI system lawful, fair and well governed? Is this AI system lawful and well documented?
Agent architecture literacy Full domain — control loops, tool surfaces, memory, orchestration topologies Introductory coverage within technical foundations Rarely addressed
Authority and identity Full domain — non-human identity, scoped grants, delegation chains, blast radius, authorisation gates Delegation and accountability at principle level Not addressed
Agentic threat modelling Full domain — injection, tool misuse, goal drift, exfiltration through action, cascading failure Failure modes mapped to legal exposure at model level Bias and privacy focus
Oversight and containment Full domain — approval gates, circuit breakers, kill switches, autonomy promotion and demotion Human oversight requirements and override principles "Human in the loop" stated as a principle
Evaluation and red teaming Full domain — trajectory evaluation, agentic red teaming, continuous regression, assurance cases Commissioning and interpreting model safety testing Limited
Evidence and forensics Full domain — trace specification, decision reconstruction, audit and litigation evidence Documentation and record-keeping obligations Documentation obligations
Delivery 4 instructor-led days · 32 contact hours Pre-course plus 2 intensive days · 28–30 hours Varies
Examination 100 scenario items · 2.5 hours · 70% pass 80 scenario items · 2 hours · 70% pass Typically 100 items · 3 hours · scaled pass
Enroll Now

What You Will Be Able to Do After AAIGP

By the end of the programme, you will be able to:

  • Interrogate an agentic system design well enough to govern it — control loops, tool surfaces, memory, orchestration — without becoming an engineer
  • Classify systems on an autonomy spectrum and identify the precise point at which an assistant or copilot crosses into agentic operation
  • Design and operate an agent authorisation gate, with defined go / no-go criteria and the evidence required for production release
  • Grant scoped, least-privilege, time-bounded authority to agents instead of standing permission, and document the rationale as governance evidence
  • Trace a delegation chain from the human principal through the agent to sub-agents and tools, and identify where accountability dilutes
  • Calculate and constrain blast radius before authorisation, not after an incident
  • Threat-model prompt injection, tool-output poisoning and exfiltration through action across every path by which untrusted content reaches agent context
  • Conduct an agentic impact assessment that treats reversibility, detectability and time-to-harm as primary variables
  • Design oversight that is meaningful at machine speed — and recognise oversight that is nominal rather than effective
  • Specify circuit breakers, safe-stop states and kill switches, and test them under realistic production conditions
  • Govern autonomy promotion and demotion against explicit evidence thresholds and automatic demotion triggers
  • Commission agentic red teaming and trajectory evaluation, interpret findings, and set remediation thresholds
  • Build a structured assurance case linking authority granted, controls applied, evidence held and residual risk accepted
  • Specify agent trace and decision logging to a standard that supports audit, regulatory scrutiny and litigation hold
  • Reconstruct a specific autonomous action end to end and explain it honestly to non-technical decision makers, regulators and claimants
  • Run agentic incident response — detection, containment, halt, evidence preservation, reversal, disclosure — when the actor is autonomous
  • Allocate liability across the agent supply chain through contractual warranties, audit rights, indemnities and insurance review
  • Scale agent governance across an enterprise portfolio without becoming the bottleneck that drives shadow agent deployment
  • Advise boards on autonomy distribution, risk appetite and the point at which authority granted exceeds assurance held

Who Should Enrol in AAIGP

AAIGP is written for people who sign off on autonomy — or who will be asked to explain it after something goes wrong.

  • Governance, risk and compliance leaders AI governance leads, risk and compliance managers, and heads of responsible AI who own the approval decision for agent deployments. You'll recognise this if: a team has asked you to approve an agent with production write access and your existing framework has no control that fits.
  • Auditors and assurance professionals Internal and external auditors, certification body personnel, and second- and third-line assurance specialists. You'll recognise this if: you have been asked to test controls over a system whose behaviour is non-deterministic and whose logs stop at the API boundary.
  • Security, identity and platform leaders Security architects, non-human identity owners, platform engineering leaders and AI engineering leads. You'll recognise this if: your agents authenticate with shared service accounts and nobody can attribute a single action to a single actor.
  • Legal, privacy and advisory professionals Legal counsel, privacy officers, and consultants advising clients on agent deployment, liability and third-party agent governance. You'll recognise this if: a client has asked who is liable when a vendor's agent takes an irreversible action on their systems, and the contract is silent.

Also strongly suited to: board members and executives setting enterprise risk appetite for autonomous action, and procurement leads assessing third-party agents and agent marketplaces.

Enroll Now

Skills You Will Gain

Authority and identity

Non-human identity governance · Credential issuance, rotation and revocation · Scoped and just-in-time authorisation · Least-privilege tool grants · Delegation-chain tracing · Confused-deputy analysis · Segregation of duties between agents · Blast-radius calculation · Agent registry and inventory design

Risk and threat modelling

Agentic risk taxonomy · Goal drift and specification gaming · Direct, indirect and cross-domain prompt injection · Tool-output poisoning · Exfiltration through action · Cross-tenant contamination analysis · Agentic impact assessment (ISO/IEC 42005 style) · Risk appetite for autonomy · Portfolio risk aggregation

Oversight, assurance and evaluation

Human-in / on / out-of-the-loop design · EU AI Act Article 14 application · Approval-gate placement and approval-fatigue controls · Circuit breakers and kill-switch testing · Graceful degradation and safe-stop states · Trajectory and behavioural evaluation · Agentic red teaming · Regression control over agent behaviour · Structured assurance cases

Evidence, response and scale

Agent trace specification · Decision reconstruction · Log integrity and litigation hold · Production monitoring for agent fleets · Agentic incident response · Multi-jurisdiction breach notification · Liability allocation across the agent supply chain · Third-party agent due diligence · Tiered enterprise governance · Board reporting on autonomy

AAIGP Curriculum — Eight Examination Domains

The examination is built directly against the domains below. Every item is written to a published performance indicator, so what you study is exactly what is assessed.
Domain I
Agentic AI Systems and Architecture Literacy · 12% · ~11 questions
  • Read and interrogate an agentic system design well enough to govern it — control loops, tool surfaces, memory and orchestration topologies at the level of detail a governance decision actually requires, without becoming an engineer.
  • I.A Distinguish agentic systems from conventional and generative AI — agency through goal direction, planning, tool use, memory and persistence; classification on an autonomy spectrum for registry and tiering
  • I.B Interpret agent architectures and control loops — planner/executor patterns, reasoning-and-action loops, reflection stages, loop termination, step budgets, recursion limits, and surfacing undeclared autonomy in a design document
  • I.C Govern tool use and the action surface — tool inventory, read vs write vs transactional vs irreversible actions, integration protocol exposure including Model Context Protocol style integrations, shadow tool access
  • I.D Govern memory, state and context — session context vs scratchpad vs long-term memory vs shared stores, retention and deletion of agent memory, context contamination, cross-tenant leakage, provenance of retrieved context
  • I.E Assess orchestration and multi-agent topologies — supervisor/worker, peer-to-peer, swarm and pipeline patterns, emergent behaviour, agent-to-agent trust boundaries, mapping topology to accountability
  • Establish what an agent is, what it may do, on whose authority, and how that authority is granted, constrained, evidenced and withdrawn. This is the domain that separates agentic governance from all prior AI governance practice.
  • II.A Establish agent identity and non-human credential governance — attributable, non-shared identities; issuance, rotation and revocation; impersonation risk; registry entries carrying owner, risk tier, autonomy level and oversight mode
  • II.B Design scoped authorisation and least-privilege tool grants — allow-lists and deny-lists, environment scoping, just-in-time and time-bounded authority, separation of development, staging and production
  • II.C Govern delegation chains and human accountability — tracing authority through sub-agents and tools, accountability dilution across hops, a named accountable owner for every production agent, three lines of defence, the confused deputy problem
  • II.D Set operational limits and blast-radius controls — spend, rate, volume and transaction limits; irreversibility thresholds that force human approval; segregation of duties between agents; data egress limits
  • II.E Operate an agent authorisation gate — go / no-go criteria, staged autonomy promotion with graduation criteria, change control over authority scope, and clean decommissioning including credentials, memory and data
  • Identify, analyse and treat the risks specific to systems that act, including adversarial risks that simply do not exist for non-agentic AI.
  • III.A Apply an agentic risk taxonomy — goal misspecification, goal drift, reward hacking, specification gaming, unbounded loops, runaway cost, unsafe tool invocation, cascading and correlated failure
  • III.B Threat-model prompt injection and adversarial input — direct, indirect and cross-domain injection; every path by which untrusted content reaches agent context; trust boundaries, provenance and input isolation
  • III.C Assess exfiltration and confidentiality risk created by action — exfiltration channels opened by tool and network access, agent-mediated aggregation of individually permitted but collectively sensitive data, outbound communication governance
  • III.D Conduct agentic impact assessment — extending ISO/IEC 42005 style assessment to autonomous action, impact on third parties who never interact with the agent, proportionality between autonomy granted and assurance required
  • III.E Integrate agentic risk into enterprise risk management — agent-specific register entries, escalation triggers, explicit risk appetite for autonomy, portfolio aggregation for board reporting
  • Design oversight that is meaningful at machine speed and machine scale — approval gates, intervention, circuit breakers, shutdown, and the human operating model around them.
  • IV.A Design meaningful human oversight — human in / on / out of the loop, EU AI Act Article 14 applied to agentic deployments, identifying nominal oversight, reviewer capacity, decision latency and automation bias
  • IV.B Implement approval gates and intervention points — gates placed at irreversible and high-consequence actions rather than uniformly, interfaces that give a reviewer enough context to refuse, timeouts and fail-safe defaults, approval-fatigue controls
  • IV.C Design containment, circuit breakers and kill switches — automated triggers on anomaly, spend, error rate and unexpected tool use; graceful degradation; who may halt an agent, by what mechanism, within what time; partial containment where full shutdown is not commercially viable
  • IV.D Manage the human and agent operating model — operator roles, rosters and escalation paths; deskilling, over-trust and vigilance decay; competence requirements for anyone supervising an autonomous system
  • IV.E Govern autonomy promotion and demotion — evidence thresholds to increase autonomy, automatic demotion triggers, pilot, shadow-mode and canary patterns, scheduled autonomy review rather than review only after failure
  • Commission, interpret and act on evidence that an agent behaves acceptably — both before authorisation and continuously in production.
  • V.A Commission behavioural evaluation — model benchmarks vs agent task and trajectory evaluation, acceptance criteria for behaviour rather than output quality alone, tool-use correctness, action sequencing, refusal behaviour, sandboxed simulation
  • V.B Direct agentic red teaming — scoping for injection, privilege escalation, exfiltration and unsafe action; rules of engagement against production-like systems; remediation thresholds; mandatory re-testing before autonomy promotion
  • V.C Apply continuous evaluation and regression control — detecting drift from model updates, prompt changes, tool changes and data changes; vendor model version changes as formal change control; retaining evaluation evidence
  • V.D Build the assurance case — a structured argument linking authority granted, controls applied, evidence held and residual risk accepted; mapping to ISO/IEC 42001 clauses and the nine GAICC control domains; identifying where assurance is asserted rather than evidenced
  • Make agent behaviour explainable and reconstructable after the fact, to a standard that satisfies auditors, regulators and courts. This is the operational meaning of governing AI you can prove.
  • VI.A Specify agent trace and decision logging — minimum trace content across inputs, retrieved context, reasoning artefacts, tool calls and parameters, results, approvals and outcomes; retention aligned to limitation periods; log integrity against tampering by humans and agents
  • VI.B Reconstruct agent decisions — end-to-end reconstruction from trace evidence, causal contribution of context vs tool output vs model behaviour, handling non-determinism honestly, distinguishing explanation of mechanism from attribution of intent
  • VI.C Monitor agents in production — operational and governance indicators for agent fleets, anomaly and cost-escalation monitoring, dashboards for first, second and third line audiences, named ownership for every alert
  • VI.D Produce audit and regulatory evidence — mapping to ISO/IEC 42001 and applicable regulation, evidence packs for internal audit and certification bodies, discovery and litigation hold over agent traces, stating evidence gaps honestly
  • Respond when an autonomous system causes harm — and allocate responsibility across the agent supply chain before that happens rather than after.
  • VII.A Execute agentic incident response — detection, containment, halt, evidence preservation, remediation and disclosure when the actor is autonomous; preserving trace evidence before remediation destroys it; determining the full scope of action already taken; reversal and compensation
  • VII.B Manage regulatory notification and disclosure — serious incident reporting applied to agentic failure, personal data breach triggers arising from agent action, multi-jurisdiction timelines, documenting and justifying decisions not to notify
  • VII.C Allocate liability across the agent supply chain — responsibilities of model provider, tool and platform provider, agent builder, orchestrator and deployer; where agentic use converts a deployer into a provider under the EU AI Act; contractual allocation, warranties, audit rights, indemnities and insurance gaps
  • VII.D Govern third-party agents and agent marketplaces — due diligence before authority is granted, agents acting on your systems under a vendor's control, transitive risk from tools an agent can discover or chain to, exit and data return requirements
  • VII.E Learn from incidents — post-incident review focused on authority and oversight design rather than operator blame, feeding findings into the risk register and autonomy levels, reporting systemic findings to the board
  • Apply current law and standards to systems they were not written for, and scale agent governance from a single pilot to an enterprise portfolio without creating a bottleneck that drives shadow deployment.
  • VIII.A Apply existing AI regulation to agentic systems — EU AI Act risk classification, GPAI obligations and Article 14; sectoral regulation in finance, health, employment and critical infrastructure; automated decision-making rights under GDPR, PIPL, PDPA, LGPD and equivalents; identifying clearly where regulation is silent
  • VIII.B Apply management system standards to agentic operation — ISO/IEC 42001 clauses and Annex A controls, the nine GAICC control domains, NIST AI RMF Govern / Map / Measure / Manage, ISO/IEC 42005 impact assessment, extending existing AIMS scope without rebuilding it
  • VIII.C Track and anticipate agent-specific regulatory development — monitoring emerging guidance and enforcement signals, assessing proposals before they bind, precautionary sequencing, advising leadership on defensible positions, avoiding over-claimed compliance
  • VIII.D Scale agent governance across the enterprise — agent registry and portfolio view, tiered governance proportionate to autonomy and impact, the GAICC Maturity Model, the seven phases of the GAICC Implementation Roadmap, designing governance that does not drive shadow agent deployment
  • VIII.E Advise boards and executives — portfolio risk, autonomy distribution and incident trend reporting; board-level risk appetite for autonomous action; the business case for agent governance investment; the journey from AI-enabled to AI-first to AI-native
Enroll Now

The Four-Day Programme Structure

  1. Day 01

    Authority and Architecture

    Domains I · II 8 Hours

    Agent architecture literacy, action surfaces, memory and orchestration; then agent identity, scoped authorisation, delegation chains, blast-radius limits and the authorisation gate.

  2. Day 02

    Risk and Control

    Domains III · IV 8 Hours

    Agentic risk taxonomy, injection and exfiltration threat modelling, agentic impact assessment; then meaningful oversight, approval gates, circuit breakers, kill switches and autonomy promotion.

  3. Day 03

    Assurance and Evidence

    Domains V · VI 8 Hours

    Behavioural and trajectory evaluation, agentic red teaming, continuous regression and assurance cases; then trace specification, decision reconstruction, production monitoring and audit evidence.

  4. Day 04

    Response, Liability and Scale

    Domains VII · VIII 8 Hours

    Agentic incident response, notification, liability allocation and third-party agent governance; then regulatory application, standards mapping, portfolio scaling and board advisory. Closes with the full mock examination and review.

Total Duration

32 Hours

Each day covers two examination domains and closes with applied scenario work and a module quiz.

Certification Benefits

Proof of a competence the market cannot yet find

AAIGP certifies the specific ability to authorise and constrain autonomy — a capability most AI governance credentials do not assess at all.

Built to ISO/IEC 17024

Issued under the international standard for personnel certification bodies, with certification decisions taken independently of training delivery and independently of the scoring platform.

Verifiable digital credential

A digital badge and verifiable certificate for LinkedIn, your CV and professional profiles, plus listing in the GAICC global credentials register.

Assessed on judgement, not vocabulary

Every item places you in a decision seat: authorise or refuse, escalate or contain, disclose or hold, promote or demote autonomy. Passing means you can make the call.

Directly applicable from day one

The programme outputs — authorisation gates, blast-radius limits, trace specifications, assurance cases — are artefacts you can take straight into your organisation.

Part of a coherent credential pathway

AAIGP sits alongside CPAIG and the GAICC ISO/IEC 42001 credentials, mapping to the same Framework, Maturity Model and Implementation Roadmap your organisation may already use.

Career Opportunities for AAIGP Holders

Agentic deployment is creating roles that did not exist two years ago, and reshaping ones that did. AAIGP positions you for:
  • Head of AI Governance / Chief AI Officer — owning enterprise risk appetite for autonomous action
  • Agentic AI Risk Lead / Autonomous Systems Risk Manager — running the authorisation gate and the agent portfolio view
  • AI Assurance and Audit Lead — testing controls over non-deterministic systems and building assurance cases
  • Non-Human Identity / Agent Security Architect — owning agent credentials, scoped grants and blast-radius design
  • AI Compliance Manager (Agentic Systems) — applying the EU AI Act, ISO/IEC 42001 and sectoral regulation to systems that act
  • AI Governance Consultant / Advisor — advising clients on agent deployment, liability allocation and third-party agent governance
  • Legal and Privacy Counsel, Autonomous Systems — drafting contractual allocation, warranties, audit rights and indemnities for agentic action
  • Board Advisor on AI Risk — escalating clearly where authority granted exceeds assurance held
Enroll Now

Your Certification Journey — Five Steps

01

Check your eligibility

Confirm you meet one of the three education-and-experience pathways, and that your relevant experience falls within the last eight consecutive years.

02

Complete 32 hours of AAIGP training

Choose any one of the three recognised routes — GAICC self-paced, a GAICC-authorised instructor or training partner, or a GAICC-recognised institution. All three are equivalent for eligibility.

03

Book your examination

Book through the GAICC hub at hub.gaicc.org. Run the system check well before exam day — an unresolved network or camera problem is the most common avoidable delay.

04

Sit the AI-proctored examination

100 scenario-based items in 150 minutes, closed book, from a quiet private space anywhere in the world. Your provisional result and domain-level feedback appear on screen at completion.

05

Receive your certification decision

The GAICC Certification Decision Authority reviews the provisional result and confirms the decision within 72 hours. Successful candidates receive a verifiable certificate and digital badge.

Enroll Now

AAIGP Examination Details

Exam at a glance

Exam Details Information
Total questions 100 scenario-based items (88 scored + 12 unscored pilot)
Format Scenario-based multiple choice, four options (A–D), single best answer
Duration 150 minutes (2 hours 30 minutes), no scheduled break
Pass mark 70% of scored questions — a minimum of 62 of 88
Scoring Scaled 100–500; 350 is the passing threshold, anchored via a modified-Angoff study
Negative marking None — answer every question
Cognitive split Understanding 15% · Application 45% · Analysis and Evaluation 40%
Reference material Closed book. No external reference material permitted
Delivery AI-proctored online on the GAICC hub (hub.gaicc.org), global browser access
Language English (other languages on request)
Results Provisional on screen at completion, with domain-level performance feedback
Certification decision Confirmed by the Certification Decision Authority within 72 hours
Enroll Now

Eligibility and Prerequisites

AAIGP is an advanced credential. The GAICC Certified Professional in AI Governance (CPAIG), or an equivalent AI governance credential, is strongly recommended but is not a formal prerequisite. The programme assumes working knowledge of AI governance fundamentals and does not re-teach them — candidates without that background should plan accordingly.

Education and experience pathways

Meet any one of the following:
Educational Background Professional Experience Required Training Requirement
Secondary qualification (high school diploma, associate degree or global equivalent) 5 years (60 months) in AI governance, risk, compliance, security, audit, data or AI delivery 32 hours of AAIGP training via any recognised route
Bachelor's degree (or global equivalent) 3 years (36 months) in AI governance, risk, compliance, security, audit, data, technology or law 32 hours of AAIGP training via any recognised route
Master's degree or professional qualification 2 years (24 months) in AI governance, law, risk, security or AI delivery 32 hours of AAIGP training via any recognised route

The 32-hour training requirement — three recognised routes

All three routes are equivalent for eligibility purposes. No route, including GAICC’s own, confers any advantage in the examination.
Route Description
1. GAICC self-paced course The 32-hour AAIGP course completed in your own time on the GAICC hub
2. GAICC-authorised Certified Instructor or Authorised Training Partner Instructor-led delivery of the 32-hour programme by an authorised partner or an authorised GAICC Certified Instructor
3. A GAICC-recognised institution 32 hours of AAIGP training delivered by an institution recognised by GAICC for this purpose
Candidates sit the examination once the training requirement is complete.
Experience recency: All relevant experience must have been gained within the last eight (8) consecutive years before you submit your AAIGP application. Experience with agentic or automated decision systems, non-human identity, security assurance or audit is recognised alongside classic AI governance experience.
Enroll Now
4.8 / 5.0 Rating

Trusted by Professionals Across AI, GRC & Legal Industries

Emma Williams

⭐⭐⭐⭐⭐

The AAIGP course gave me a practical understanding of how AI agents should be governed when they start taking real-world actions. The scenarios and governance exercises made the concepts easy to apply in my role.

Sophie Laurent

⭐⭐⭐⭐⭐

This certification gave me a structured approach to managing agentic AI risk. I especially appreciated the focus on agent authority, identity, human oversight and containment. It is highly relevant for organisations adopting autonomous AI.

Lucas Ferreira

⭐⭐⭐⭐⭐

AAIGP is one of the most practical AI governance certifications I’ve completed. The modules on agentic risk, red teaming, audit evidence and incident response have strengthened the way I approach AI governance.

Isabella Turner

⭐⭐⭐⭐⭐

The course strikes the right balance between technical understanding and governance decision-making. I particularly valued the practical focus on authorisation, accountability and controlling AI agents in real-world environments.

Delivery Format and What's Included

How the programme is delivered

The AAIGP programme is 32 contact hours, structured as four eight-hour days covering two examination domains each. You can complete those hours instructor-led — over four consecutive days or a schedule set by your training provider — or self-paced on the GAICC hub, whichever suits your role and time zone.

Instructor-led delivery closes each day with applied scenario work and a module quiz, and Day 4 finishes with a full mock examination and review.

What's included

  • 32 contact hours of AAIGP programme content across four structured days
  • Applied scenario work and end-of-day module quizzes
  • Full mock examination and review
  • Domain-by-domain preparation mapped to the published Examination Content Outline
  • Digital certificate and verifiable badge on successful certification
  • Listing in the GAICC global credentials register
  • 60 CPD hours' worth of eligible professional learning toward your certification cycle. CONFIRM the exact CPD/PDU award figure for AAIGP training completion with the certification team
  • Exam-day requirements
  • A quiet, private space and a stable internet connection
  • A working camera and microphone — run the system check in advance
  • No notes, devices or assistance (closed book) unless an approved adjustment applies

Member Price

US$898

Full Price

US$1075

Book Your Seat Now

US$99 is included in the above member price.

Renewal and re-examination fees – Published annually

Maintaining Your AAIGP Certification

AAIGP is valid for three years from the date of certification. To maintain it, you earn 60 CPD hours per three-year cycle and reaffirm the GAICC Code of Ethics.
CPD Category What Counts Minimum Hours
1. Professional learning GAICC-recognised training, conferences, webinars or workshops in AI governance, agentic AI, security, law, regulation or assurance 20
2. Practical application Direct involvement in agent authorisation, agentic risk assessment, oversight design, agent evaluation, audit or incident response 20
3. Contribution and knowledge sharing Publishing, mentoring, research, contributing to AI governance or agent safety standards, policy papers or community initiatives 10
4. Elective activities Additional learning or engagement supporting continuous professional growth in AI-adjacent disciplines 10
Total per three-year cycle 60 CPD hours

At least 20 of those hours must come from practical application — this credential is maintained by doing the work, not only by attending sessions.

Where CPD evidence is insufficient, where certification has lapsed beyond any grace period, or where the standard or scheme changes materially, re-examination is required. Full rules are set out in the published recertification policy for this certification.

Use of the certificate and GAICC marks: certified persons may state that they hold the certification while it is valid and use the GAICC certification mark in accordance with the Use of Certificates and Marks procedure. Certification applies to the named individual, is not transferable, and must not be used misleadingly or to imply accreditation that has not been granted. Lapsed or withdrawn certificants must cease use of the certification and the marks.

Enroll Now

The GAICC Code of Professional Conduct

Every AAIGP candidate and certified professional agrees to the GAICC Code of Ethics. AAIGP adds one obligation specific to autonomy, and it is the one that defines the credential.

2 · Proportionate Authority

Never recommend or approve a level of agent autonomy that exceeds the assurance evidence held. Where authority and evidence diverge, say so in writing.
Principle Obligation
1. Integrity and Fairness Act honestly in all professional activities without bias, conflict of interest or misrepresentation of competence. Disclose potential conflicts proactively.
2. Proportionate Authority Never recommend or approve a level of agent autonomy that exceeds the assurance evidence held. Where authority and evidence diverge, say so in writing.
3. Human Rights and Wellbeing Ensure autonomous systems under your governance influence respect human dignity, fairness, privacy and non-discrimination — including for people who never interact with the system.
4. Transparency and Accountability Promote explainable and reconstructable agent behaviour. Ensure a named human remains accountable for every autonomous action.
5. Data Protection and Privacy Uphold confidentiality and data-protection principles across agent memory, tool access and trace data, consistent with applicable laws and international standards.
6. Professional Competence Maintain current knowledge in a field that is changing faster than its regulation, through continuous learning and active engagement with evolving standards.
7. Reporting and Mitigation of Misuse Take appropriate action when encountering unsafe autonomy, suppressed incidents, disabled controls or violations of applicable regulation or professional standards.
8. Global Responsibility Recognise that autonomous systems act across borders. Apply governance standards that protect individuals regardless of jurisdiction.
Breaches may lead to withdrawal of certification.
Have Questions?

Frequently asked question.

AAIGP is an advanced practitioner credential for professionals who authorise, oversee, assure and shut down AI systems that take action. It governs autonomy, not just output — covering agent architecture literacy, authority and identity, agentic risk, oversight and containment, evaluation and assurance, evidence and forensics, incident response and liability, and enterprise scaling.
CPAIG governs AI systems broadly across regulation, ethics and lifecycle, with agentic AI as one domain of eight. AAIGP is entirely about systems that act: authority, identity, delegation, containment, evaluation, evidence and liability. If CPAIG asks whether an AI system is lawful and well governed, AAIGP asks what it is permitted to do and how that permission is bounded and evidenced.
No. CPAIG or an equivalent AI governance credential is strongly recommended but is not a formal prerequisite. AAIGP assumes working knowledge of AI governance fundamentals and does not re-teach them, so candidates without that background should expect a steeper climb.
AI governance leads, risk and compliance managers, internal and external auditors, security and identity architects, platform and AI engineering leaders, legal and privacy professionals, and consultants advising on agent deployment.
100 scenario-based multiple-choice questions with four options each, completed in 150 minutes. The exam is closed book and AI-proctored online on the GAICC hub at hub.gaicc.org. The pass mark is 70% of the 88 scored questions.
Because the competence being certified is judgement, not recall. Every item places you in a decision seat — authorise or refuse, escalate or contain, disclose or hold, promote or demote autonomy. No question tests definition recall alone, and no question can be answered from vendor terminology.
No. There is no penalty for a wrong answer, so you should answer every question.
32 contact hours, delivered as four instructor-led days of eight hours each, or completed self-paced on the GAICC hub.
Whichever suits your schedule. GAICC self-paced, a GAICC-authorised Certified Instructor or Authorised Training Partner, and a GAICC-recognised institution are all equivalent for eligibility, and no route — including GAICC’s own — confers any advantage in the examination.
A provisional result with domain-level feedback is displayed on screen the moment you finish. The Certification Decision Authority reviews it and confirms the certification decision within 72 hours.
Domain-level feedback is provided with every result so revision before a re-sit can be targeted. Re-sit arrangements and any applicable fee are published at certifications.gaicc.org.
Three years. Renewal requires 60 CPD hours per cycle — including at least 20 hours of practical application — plus a renewal application, reaffirmation of the GAICC Code of Ethics, and the applicable fee.
A verifiable digital certificate and a digital badge for LinkedIn, CVs and professional profiles, plus listing in the GAICC global credentials register.
Yes. Request a reasonable adjustment at booking with supporting information. GAICC accommodates legitimate needs while preserving examination security and assessment integrity.
AAIGP operationalises the GAICC AI Governance Framework for hybrid human and autonomous-agent teams, and maps directly to the Maturity Model dimensions and the Implementation Roadmap phases.
The credential is issued by the Global AI Certification Council under ISO/IEC 17024, the international standard for bodies certifying persons, and is delivered and examined globally. It applies frameworks including the EU AI Act, ISO/IEC 42001, ISO/IEC 42005, ISO/IEC 23894, the NIST AI RMF and the OECD AI Principles. CONFIRM any additional accreditation claim (e.g. CPDSO) with the certification team before publishing.
Instructor

Dr Faiz Rasool

Director at the Global AI Certification Council (GAICC) and PM Training School

Linkedin-in

A globally certified instructor in ISO/IEC, PMI®, TOGAF®, SAFe®, and Scrum.org disciplines. With over three years’ hands-on experience in ISO/IEC 42001 AI governance, he delivers training and consulting across New Zealand, Australia, Malaysia, the Philippines, and the UAE, combining high-end credentials with practical, real-world expertise and global reach.

Other GAICC ISO/IEC 42001 Courses

Best Selling

GAICC ISO/IEC 42001 Lead Implementer Certification

  • Updated: April, 2026
  • 4 Days
⭐ 4.8
29,899+
Best Selling

GAICC ISO/IEC 42001 Lead Auditor Certification

  • Updated: April, 2026
  • 4 - 5 Days
⭐ 4.8
23,357+
Best Selling

GAICC ISO/IEC 42001 Internal Auditor Certification

  • Updated: April, 2026
  • 2 - 3 Days
⭐ 4.8
25,183+

Start Your ' Agentic AI Governance Professional Certification (AAIGP') Today

4.8 / 5.0 Rating
  • Become an AI governance, technical risk architecture, lifecycle governance, and strategic board advisory.
  • Land High Demand AI Governance Roles
  • Average of $18k salary hike

Self-Paced Course (Certification Exam included)

Member Price

US$898

Full Price

US$1075

GAICC Annual Membership US$99 is included in the above member price.

Enroll Now
Trusted by 24,000+ Professionals

Ready to Become an Agentic AI Governance Professional Certification ?

Start Your Certification Now
AI Governance Job Insights 2026
AI Governance Job Insights 2026

AI Governance Job Insights 2026