Exampractice

Certification path

AI certification path

AI certifications are not one ladder but several, because each cloud and tool vendor runs its own. Most people start with a foundational exam from the platform they already use — AWS AI Practitioner, Azure AI Fundamentals or Google's Generative AI Leader — then move to an engineering credential on the same platform. Two newer tracks sit beside those: generative-AI and agent certifications from NVIDIA, Databricks, Microsoft and Salesforce, and governance credentials for people who run AI programmes rather than build them.

Where to start

If you work on AWS, start with AI Practitioner and go on to Machine Learning Engineer – Associate. On Azure, Azure AI Fundamentals then AI Engineer Associate. If you manage AI rather than build it, the Google Generative AI Leader or PMI's CPMAI is the right first exam, and AIGP is the one for governance and compliance roles.

  1. 01

    Beginner AI

    4 exams

    No prerequisites. Vocabulary, the main services and responsible-AI basics; enough to work alongside AI projects.

  2. 02

    AI engineering

    3 exams

    Building, training and deploying models on a platform. Hands-on experience expected.

  3. 03

    AI agents

    1 exam

    The newest track: building and running agents that act, not just answer.

  4. 04

    AI governance and management

    2 exams

    For people who run AI programmes: risk, law, procurement and delivery rather than code.

Work the whole ladder on one plan

$39.99 a month covers all 10 exams on this page and every other exam on the site — which is the point of a path. Or read 10 questions from any of them first, free.

See pricing

Common questions

Where should I start on the AI certification path?
If you work on AWS, start with AI Practitioner and go on to Machine Learning Engineer – Associate. On Azure, Azure AI Fundamentals then AI Engineer Associate. If you manage AI rather than build it, the Google Generative AI Leader or PMI's CPMAI is the right first exam, and AIGP is the one for governance and compliance roles.
How many AI exams are there?
10 of them are covered here, arranged into 4 tiers: beginner AI, AI engineering, AI agents, AI governance and management. Retired versions still in our catalogue are deliberately left off — the point of this page is what to sit next.
Do I have to take the AI exams in order?
AI certifications are not one ladder but several, because each cloud and tool vendor runs its own. Most people start with a foundational exam from the platform they already use — AWS AI Practitioner, Azure AI Fundamentals or Google's Generative AI Leader — then move to an engineering credential on the same platform. Two newer tracks sit beside those: generative-AI and agent certifications from NVIDIA, Databricks, Microsoft and Salesforce, and governance credentials for people who run AI programmes rather than build them.
Can I practise these before paying?
Yes — 10 free questions on every exam on this page, with no card and no account. Each also has a timed practice test of 65 questions in 130 minutes.
What does it cost to prepare for the whole path?
$39.99 a month opens every exam on this page and every other exam on the site, which is the plan that makes sense if you are working through a ladder. A single exam on its own is a one-time $19.99 for 30 days.

Other certification paths