Artificial Intelligence Insights

A Professional's Guide to AI Certification Training

Twelve AI certifications from CompTIA, AWS, PMI, ISACA, EC-Council, Google Cloud, Microsoft, IAPP and NVIDIA, compared side by side. Exam format, cost, prerequisites and renewal, checked against each certifying body, plus which one fits your role.

The quick answer

Start with the work you do now. If you're new to AI, AWS Certified AI Practitioner or Microsoft's AI-901 is the place to begin. Engineers who build models pick by cloud: AWS Machine Learning Engineer, Google Professional Machine Learning Engineer or Microsoft AI-103. Security people look at CompTIA SecAI+ first and ISACA's AAISM once they hold CISM or CISSP. Auditors go to AAIA. IAPP's AIGP is the governance pick. And if you manage AI projects or programs, compare PMI-CPMAI with EC-Council's CAIPM.

  • 12AI certifications we train for, from 9 certifying bodies
  • $100 to $799Standard U.S. exam prices we could confirm, AWS AI Practitioner lowest and AIGP for non-members highest
  • 2Exams that require another certification first, AAIA and AAISM
  • Sep 28, 2026Last day AWS offered MLA-C01 in English. MLA-C02 is in beta now

AI certifications have changed a lot in a short time. In the last year Microsoft replaced both of its Azure AI exams, AWS put a new Machine Learning Engineer exam into beta, CompTIA launched SecAI+, ISACA added two advanced AI credentials and PMI rebuilt CPMAI under its own name. A lot of the comparison charts online still show the old versions.

So I went to each certifying body's own pages, exam guides and candidate handbooks and checked every figure in this guide on October 11, 2026. Where a body doesn't publish a number, like Microsoft's question counts or Google's passing score, the table says so instead of guessing.

All 12 AI certifications at a glance

The 12 exams compared, from each certifying body's pages checked October 11, 2026. "Not listed" means the body doesn't publish that figure. Swipe sideways on a phone to see every column.
CertificationLevelQuestionsTimePassing scoreU.S. exam priceStays valid by
CompTIA SecAI+ (CY0-001)Intermediate securityUp to 60, multiple choice and performance-based60 minutes600 on a 100 to 900 scaleNot listed on the exam pageCompTIA CE program, $75 fee per 3-year cycle
AWS Certified AI Practitioner (AIF-C01)Foundational65, 50 of them scored90 minutes700 on a 100 to 1,000 scale$100Valid 3 years, recertify by exam
AWS Certified Machine Learning Engineer, AssociateAssociate85 on the MLA-C02 beta, 65 on MLA-C01170 minutes beta, 130 minutes MLA-C01720 on a 100 to 1,000 scale$75 beta, $150 standardValid 3 years, recertify by exam
Google Cloud Professional Machine Learning EngineerProfessional50 to 60, multiple choice and multiple select2 hoursNot listed$200 plus taxRenewal exam inside Google's eligibility window
Microsoft Azure AI Fundamentals (AI-901)FundamentalsNot listedNot listed700Varies by countryNot covered here
Microsoft Azure AI Apps and Agents Developer Associate (AI-103)AssociateNot listed120 minutes700Varies by countryExpires yearly, free online renewal assessment
NVIDIA Certified Associate: Generative AI LLMs (NCA-GENL)Associate50 to 60, multiple choice1 hourNot listed$125Valid 2 years, retake to recertify
ISACA AAISMAdvanced, needs CISM or CISSP90, multiple choice150 minutes450 on a 200 to 800 scale$459 ISACA members, $599 non-members10 CPE a year, 30 in 3 years, $20 or $35 a year
ISACA AAIAAdvanced, needs CISA or a listed designation90, multiple choice150 minutes450 on a 200 to 800 scale$459 ISACA members, $599 non-members10 AI CPE hours a year
IAPP AIGPProfessional governance100, multiple choice with case studies2.75 hours with a 15-minute break300 on a 100 to 500 scale$649 IAPP members, $799 non-members20 CPE credits per 2-year term
PMI-CPMAINo experience required120, 20 of them unscored pretest160 minutesNot listedSold as a bundle with PMI's prep course30 PDUs every 3 years
EC-Council C|AIPM (312-41)Program leadership100, multiple choice3 hours70% to 80%, set per exam formNot listedNot covered here

Two columns tell you the most. Prerequisites decide whether you can even sit the exam, and the ISACA pair are the only two here that check. Renewal tells you what you're signing up for after you pass. Microsoft's associate certifications expire every year, while most of the others run two or three years.

New, beta and recently replaced exams

Several of these exams changed in the last year. Study from old material and you can end up preparing for an exam that no longer exists.

  • September 30, 2025PMI introduced PMI-CPMAI, replacing Cognitive Project Management in AI (CPMAI) v7.
  • February 17, 2026CompTIA launched SecAI+ (CY0-001), the first exam in its expansion series. CompTIA estimates V1 retires about 3 years after launch.
  • April 2026Microsoft opened AI-103 and AI-901 as beta exams. AI-103 covers agents and generative AI in Microsoft Foundry, and AI-901 became the new route to Azure AI Fundamentals.
  • June 30, 2026Microsoft retired AI-102 and AI-900. Since then AI-901 is the only exam for Azure AI Fundamentals, and AI-103 has replaced AI-102. Neither exam page carries a beta label now, and the AI-901 page lists no retirement date.
  • September 28, 2026Last day AWS offered MLA-C01 in English. MLA-C01 stays open in Japanese, Korean and Simplified Chinese until MLA-C02 is generally available.
  • September 29, 2026 to nowMLA-C02 beta delivery is running in English only. AWS hasn't set a general availability date yet.

About the MLA-C02 beta. Beta exams carry extra unscored items, which is why it's 85 questions in 170 minutes. AWS says beta results usually come within 5 business days. The four domains are the same as MLA-C01, with generative AI, Amazon Bedrock, RAG, agents and responsible AI added to the tasks.

ISACA's AAIA and AAISM are new as well. They're advanced credentials that sit on top of an existing ISACA or ISC2 certification, which is a different model from every other exam here. EC-Council's C|AIPM is also new, built for people who run AI programs rather than build models.

Who each one fits

First AI cert

AWS AI Practitioner or AI-901

Both are entry points. AWS aims AI Practitioner at people who use AI on AWS but don't necessarily build it, which is why business analysts, product people and managers take it. AI-901 is more hands-on. Microsoft expects some Python and familiarity with Azure resources.

Builders

MLA, Google PMLE, AI-103, NVIDIA

For developers and ML engineers. Pick the cloud you work in. NVIDIA's associate exam fits people who work with LLMs directly, and its learning path uses NVIDIA tools.

Security

SecAI+ and AAISM

SecAI+ is for working security practitioners who now have to defend AI systems and use AI in their own operations. AAISM is the management layer for people who already hold CISM or CISSP.

Audit and governance

AAIA and AIGP

AAIA is for IT auditors and advisors with CISA or a qualifying accounting or audit designation. AIGP is IAPP's governance credential, a natural fit for privacy, legal, compliance and risk teams.

Managing AI work

PMI-CPMAI and C|AIPM

CPMAI teaches PMI's six-phase method for running AI projects and asks for no prior experience. C|AIPM is EC-Council's program-level credential, focused on strategy, adoption, governance and ROI. EC-Council says no coding background is needed.

The 12 certifications, one by one

Each card covers what the exam tests, the checked facts and how our course is set up. Click a badge in the table above to jump to its card.

Foundations

AWS Certified AI Practitioner badge

AWS Certified AI Practitioner

Foundational knowledge of AI, ML and generative AI on AWS. Domains run from AI and ML fundamentals to responsible AI and security. The biggest one is Applications of Foundation Models, at 28 percent.

  • AIF-C01, 65 questions in 90 minutes, $100
  • 700 to pass, valid 3 years
  • AWS suggests up to 6 months of exposure to AI and ML on AWS
  • Our course: 10 modules, most students finish in 3 to 6 weeks part-time
Microsoft Certified Azure AI Fundamentals badge

Microsoft Azure AI Fundamentals (AI-901)

The new exam behind Azure AI Fundamentals. Identifying AI concepts and capabilities is 40 to 45 percent, and building AI solutions in Microsoft Foundry is 55 to 60 percent. That second half makes it more practical than the old AI-900.

  • 700 to pass, price set by country
  • Microsoft expects Python syntax knowledge and familiarity with Azure resources
  • Our course: 8 modules mapped to Microsoft's skills measured as of April 15, 2026

Building AI solutions

MLA-C02 in beta
AWS Certified Machine Learning Engineer Associate badge

AWS Certified Machine Learning Engineer, Associate

Putting ML workloads into production on AWS: data preparation, model development, deployment and orchestration, then monitoring and security. MLA-C02 adds foundation models and agentic workflows to the same four domains.

  • MLA-C02 beta: 85 questions, 170 minutes, $75
  • MLA-C01: 65 questions, 130 minutes, $150
  • 720 to pass, valid 3 years
  • AWS recommends at least 1 year with SageMaker AI and related services
  • Our course: 6 modules built on the exam domains
Google Cloud Professional Machine Learning Engineer badge

Google Cloud Professional Machine Learning Engineer

Building, scaling, serving and monitoring traditional and generative AI models on Google Cloud. Google recently updated the exam for the move from Vertex AI to Gemini Enterprise Agent Platform, so check the current exam guide for product names.

  • 50 to 60 questions in 2 hours, $200 plus tax
  • No prerequisites. Google recommends 3 or more years in industry, 1 or more on Google Cloud
  • The exam doesn't test coding directly, but expects you to read Python and SQL
  • Our course: 10 modules with hands-on labs and case studies
Replaced AI-102
Microsoft Certified Azure AI Apps and Agents Developer Associate badge

Microsoft Azure AI Apps and Agents Developer Associate (AI-103)

Building agents and AI apps with Python and Microsoft Foundry. Generative AI and agentic solutions is the largest domain at 30 to 35 percent, followed by planning and managing an Azure AI solution at 25 to 30 percent.

  • 120 minutes, 700 to pass, price set by country
  • Interactive components may appear on the exam
  • Expires every year. Renew free with an online assessment on Microsoft Learn
  • Our course: 10 modules mapped to skills measured as of April 16, 2026
NVIDIA Certified Associate Generative AI LLMs badge

NVIDIA Certified Associate: Generative AI LLMs

An entry-level exam on developing and maintaining LLM applications. Core ML and AI knowledge is 30 percent, then software development at 24, experimentation at 22, data analysis at 14 and trustworthy AI at 10.

  • NCA-GENL, 50 to 60 questions in 1 hour, $125
  • NVIDIA asks for a basic understanding of generative AI and LLMs
  • English only, valid 2 years
  • Our course: 10 modules from LLM basics to NeMo, deployment and governance

Securing AI

New in 2026
CompTIA SecAI+ certification badge

CompTIA SecAI+

Securing AI systems is 40 percent of the exam. The rest is AI-assisted security at 24, governance, risk and compliance at 19 and basic AI concepts at 17. Expect adversarial attacks, model and pipeline protection and frameworks like the NIST AI RMF.

  • CY0-001, up to 60 questions in 60 minutes
  • 600 to pass on a 100 to 900 scale
  • CompTIA recommends 3 to 4 years in IT and 2 or more in hands-on security
  • Our course: 10 modules, most students finish in 4 to 8 weeks part-time
ISACA AAISM Advanced in AI Security Management badge

ISACA AAISM

AI security management for people who already run security programs. AI Technologies and Controls is 38 percent, and AI Governance and Program Management and AI Risk Management are 31 percent each.

  • 90 questions in 150 minutes, 450 to pass
  • You must hold an active CISM or CISSP
  • $459 for ISACA members, $599 for non-members, plus a $50 application fee after you pass
  • Our course: 10 modules, live virtual instructor-led

Audit and governance

ISACA AAIA Advanced in AI Audit badge

ISACA AAIA

Auditing AI. AI Operations carries 46 percent, which covers data management, the AI life cycle, change management, testing and AI incidents. AI Governance and Risk is 33 percent and AI Auditing Tools and Techniques is 21.

  • 90 questions in 150 minutes, 450 to pass
  • You need an active CISA, or a listed designation like CIA or CPA with an IT audit or advisory focus
  • $459 for ISACA members, $599 for non-members, plus a $50 application fee
  • Our course: 10 modules, live virtual instructor-led
IAPP AIGP Artificial Intelligence Governance Professional badge

IAPP AIGP

Responsible development, deployment and management of AI systems. The body of knowledge covers AI governance foundations, how laws, standards and frameworks apply to AI, and how to govern AI development and deployment.

  • 100 questions in 2.75 hours with a 15-minute break
  • 300 to pass on a 100 to 500 scale
  • $649 for IAPP members, $799 for non-members
  • Our course: 10 modules ending with exam prep and case studies

Managing AI projects and programs

PMI-CPMAI Certified Professional in Managing AI badge

PMI-CPMAI

Running AI projects with the six CPMAI phases, from matching AI to a business need through data, model development, testing and putting AI into operation. PMI keeps it tool-agnostic.

  • 120 questions in 160 minutes, 20 of them unscored
  • No experience required
  • PMI requires its 21-hour PMI-CPMAI Exam Prep Course before you can schedule. It comes bundled with the certification
  • Our course: 10 modules on the six phases, exam format and certification steps
EC-Council C|AIPM Certified AI Program Manager logo

EC-Council C|AIPM

Leading AI programs across an organization: readiness, use case selection, strategy and roadmaps, change management, governance and measuring value. EC-Council calls it a business-led program.

  • Exam 312-41, 100 questions in 3 hours
  • Passing score between 70 and 80 percent, depending on the exam form
  • No coding background required
  • Our course: 10 modules over three instructor-led days, in person or live online

Role paths and what to take first

Business, product or IT manager

AWS AI Practitioner

No building required and the lowest price on the list. Add CPMAI or C|AIPM if you end up owning AI projects.

Azure developer

AI-901, then AI-103

AI-901 gets you into Foundry. AI-103 is where agents and RAG get serious. Plan on yearly renewal for AI-103.

ML engineer or data scientist

MLA or Google PMLE

Pick your cloud. On AWS, AI Practitioner first is optional, and earning MLA also renews it.

Security analyst or engineer

SecAI+

CompTIA suggests Security+ level knowledge first. Our first cybersecurity certification guide covers that step.

Security manager with CISM or CISSP

AAISM

It builds on the cert you have. Still choosing between those two? Read CCISO vs CISSP vs CISM vs SecurityX.

Auditor, privacy or compliance lead

AAIA or AIGP

AAIA if you hold CISA and audit AI systems. AIGP if your work is policy, law and AI risk across the business.

A common pairing is a technical cert plus a governance or management cert. A developer with AI-103 and AIGP can build the agent and explain the controls around it. A PMP holder adding CPMAI can run the project and talk data readiness with the engineers. See what changed on the PMP exam if you're doing both.

Prerequisites and experience

Only two of these exams check anything before you sit. Everything else is a recommendation, so you can register whenever you're ready.

Required versus recommended, from each body's pages
ExamRequired to sitRecommended
AAISMActive CISM or CISSPExperience assessing, implementing and maintaining AI systems
AAIAActive CISA, or a listed designation with an IT audit or advisory focusExperience auditing or advising on AI systems
PMI-CPMAIPMI's 21-hour exam prep course, bundled with the examNo experience required
SecAI+Nothing3 to 4 years in IT, 2 or more in security
Google PMLENothing3 or more years in industry, 1 or more on Google Cloud
AWS MLANothing1 year with SageMaker AI and 1 year in a related role
AWS AI PractitionerNothingUp to 6 months of AI and ML exposure on AWS
AI-103 and AI-901NothingPython, plus familiarity with Azure
NVIDIA NCA-GENLNothingBasic understanding of generative AI and LLMs

Passing ISACA's exam isn't the last step. For AAIA and AAISM you then apply to ISACA, pay the $50 application fee and keep your qualifying certification active. You have five years from your pass date to apply.

Renewal and continuing education

What it takes to keep each credential, where the body publishes it
CredentialHow it stays active
AI-103Expires every year. Renew with a free online assessment on Microsoft Learn
AWS AI PractitionerValid 3 years. Pass the latest exam, or earn MLA, which renews it automatically
AWS MLAValid 3 years. Pass the latest version of the exam
NVIDIA NCA-GENLValid 2 years. Retake the exam
IAPP AIGP2-year term, 20 CPE credits tied to the AIGP body of knowledge. Non-members pay a $250 maintenance fee each term, members don't
PMI-CPMAI30 PDUs every 3 years
ISACA AAISM10 CPE hours a year, 30 in 3 years, $20 a year for members or $35 for non-members, and an active CISM or CISSP
ISACA AAIA10 CPE hours in AI every year, starting the calendar year after you certify
CompTIA SecAI+CompTIA's continuing education program, with a $75 fee for the three-year period

ISACA charges one annual maintenance fee across all of its advanced certifications, so holding AAIA and AAISM together doesn't double it. Our guide to keeping your certification active covers CPE, CEU and PDU rules across bodies.

How our AI courses work

We offer training for all 12. AWS AI Practitioner is on the course calendar with a new class every 2 to 3 weeks. For the other 11, email us and we'll tell you the next start. Delivery depends on the course. AAIA and AAISM run as live virtual instructor-led classes, and C|AIPM runs over three instructor-led days, in person or live online.

Two things to sort out before you enroll. If you're aiming at AAIA or AAISM, confirm with ISACA that your current certification qualifies. If you're taking an exam that just changed, like AWS MLA, tell us which version you plan to sit so we point you at the right objectives.

Since August 2022, 2,405 Accumentum students have certified out of about 2,435 attempts, a 98.8% pass rate.

Ways to pay for your course

PathPay

Pay monthly

PathPay® spreads the course cost into monthly payments. It's a payment plan, not a loan.

Employer

Have your employer pay

Many employers set aside training funds. Ask before you enroll, since many want their paperwork done before the course starts. The funding page covers the options.

Military

Army CA and Air Force COOL

Accumentum is an Army CA and Air Force COOL Approved Vendor, but the Army or the Department of the Air Force decides eligibility and payment. Get approval before you enroll. See our Army CA and Air Force COOL pages.

Want more detail on funding? Read how to pay for IT certification training. Questions about which AI course fits? Email us at info@accumentum.net.

Frequently asked questions

Which AI certification should I get first?

If you're new to AI, start with AWS Certified AI Practitioner or Microsoft AI-901. Engineers should pick the exam for their cloud, security pros usually start with CompTIA SecAI+, and project or program managers should compare PMI-CPMAI with EC-Council C|AIPM.

Which AI certifications need another certification first?

Two of them. ISACA's AAISM requires an active CISM or CISSP. ISACA's AAIA requires an active CISA or a listed designation like CIA or CPA held by someone in an IT audit or advisory role.

Is AI-103 the same as AI-102?

No. AI-103 replaced AI-102, which Microsoft retired on June 30, 2026. AI-103 leads to the Azure AI Apps and Agents Developer Associate certification and puts the most weight on generative AI and agentic solutions in Microsoft Foundry.

Is AI-901 replacing AI-900?

It already has. AI-900 retired on June 30, 2026, and AI-901 is now the exam for Microsoft Certified: Azure AI Fundamentals. It adds hands-on work in Microsoft Foundry, which is 55 to 60 percent of the exam.

Which AWS Machine Learning Engineer exam should I take?

In English, the MLA-C02 beta is the only version available since MLA-C01 ended on September 28, 2026. The beta has 85 questions in 170 minutes for $75. MLA-C01 is still offered in Japanese, Korean and Simplified Chinese until MLA-C02 is generally available.

How much do AI certification exams cost?

As of October 11, 2026, standard U.S. prices we could confirm are $100 for AWS AI Practitioner, $125 for NVIDIA NCA-GENL, $150 for AWS MLA-C01, $200 for Google Professional ML Engineer, $459 or $599 for AAIA and AAISM, and $649 or $799 for AIGP. Microsoft prices vary by country.

Do I need to code for these AI certifications?

Not for all of them. AWS AI Practitioner, PMI-CPMAI, C|AIPM and AIGP don't require coding. Microsoft expects Python for AI-901 and AI-103, and Google expects you to read Python and SQL on the Professional ML Engineer exam.

What is the difference between PMI-CPMAI and C|AIPM?

PMI-CPMAI is a project method. It teaches PMI's six CPMAI phases for running an AI project and needs no prior experience. EC-Council's C|AIPM works at the program level, covering AI strategy, adoption, governance and ROI across an organization.

Sources

All checked October 11, 2026.

  1. CompTIA, SecAI+ certification page: CY0-001 exam details, launch date, domains, recommended experience and retirement estimate: comptia.org SecAI+
  2. CompTIA, Continuing Education renewal fees, SecAI+ $75 for the three-year period: comptia.org CE renewal fees
  3. AWS, Certified AI Practitioner page: exam length, format, price, validity and recertification: aws.amazon.com AI Practitioner
  4. AWS, AIF-C01 exam guide: target candidate, scored questions, passing score and domain weights: AIF-C01 exam guide
  5. AWS, Certified Machine Learning Engineer Associate page: MLA-C01 and MLA-C02 beta details, key dates and validity: aws.amazon.com ML Engineer Associate
  6. AWS, MLA-C02 exam guide: target candidate, passing score and domains: MLA-C02 exam guide
  7. PMI, PMI-CPMAI Exam Prep Course and Certification page: exam length, no experience required, prep course and maintenance: pmi.org PMI-CPMAI
  8. PMI, PMI-CPMAI Exam Content Outline: pretest questions and the required prep course: PMI-CPMAI ECO (PDF)
  9. PMI, CPMAI to PMI-CPMAI transition FAQs, September 30, 2025 launch: pmi.org CPMAI FAQ
  10. ISACA, AAIA certification page: qualifying designations, exam fees and certification steps: isaca.org AAIA
  11. ISACA, AAIA Exam Content Outline: 90 questions and domain weights: isaca.org AAIA exam content outline
  12. ISACA, AAIA Exam Candidate Guide: exam length and the 450 passing score: AAIA Exam Candidate Guide (PDF)
  13. ISACA, How to get AAIA certified: application fee and AI CPE requirement: isaca.org get AAIA certified
  14. ISACA, AAISM certification page: prerequisite, exam fees and certification steps: isaca.org AAISM
  15. ISACA, AAISM Exam Content Outline: 90 questions and domain weights: isaca.org AAISM exam content outline
  16. ISACA, AAISM Exam Candidate Guide: exam length and the 450 passing score: AAISM Exam Candidate Guide (PDF)
  17. ISACA, Maintain AAISM: CPE hours and annual maintenance fee: isaca.org maintain AAISM
  18. EC-Council, Certified AI Program Manager course page: exam code, questions, duration, passing score and FAQs: EC-Council C|AIPM
  19. Google Cloud, Professional Machine Learning Engineer certification page: exam length, fee, format, prerequisites and recommended experience: cloud.google.com Professional ML Engineer
  20. Microsoft Learn, Azure AI Apps and Agents Developer Associate certification page: exam time and languages: learn.microsoft.com AI-103 certification
  21. Microsoft Learn, Exam AI-103 study guide: passing score, renewal and skills measured as of April 16, 2026: AI-103 study guide
  22. Microsoft Tech Community, New Azure AI Apps and Agents Developer Associate, April 16, 2026: AI-103 beta launch: techcommunity.microsoft.com
  23. Microsoft Learn, Exam AI-901: passing score, skills measured and retirement date: learn.microsoft.com AI-901
  24. Microsoft Tech Community, Evolving the Azure AI Fundamentals certification: AI-901 beta and the AI-900 retirement on June 30, 2026: techcommunity.microsoft.com
  25. IAPP, AIGP exam in the IAPP store: questions, time, member and non-member prices and maintenance: store.iapp.org AIGP exam
  26. IAPP, Certification FAQs: the 100 to 500 scale and 300 passing score: iapp.org certification FAQs
  27. IAPP, AIGP Body of Knowledge: iapp.org AIGP body of knowledge
  28. NVIDIA, Generative AI LLMs certification page: exam details, price, validity and topic weights: nvidia.com NCA-GENL

Share this article

Keep Exploring

Find your next certification advantage.

Return to Accumentum Insights or jump into related training and exam preparation.