Pick the AI certification that matches the cloud your company already runs. On AWS, take AWS Certified AI Practitioner (AIF-C01). On Google Cloud or Workspace, take Google Cloud Generative AI Leader. In a Microsoft shop, take AI-901 (Azure AI Fundamentals), because AI-900 and AI-102 both retired on June 30, 2026. Move to AI-103 only if you’re starting to build AI apps in Python.
Quick comparison
| Credential | Level | Exam facts (official page) | Best for |
|---|---|---|---|
| AWS Certified AI Practitioner | Foundational | 90 minutes, 65 questions, 100 USD | AWS shops, broad AI literacy |
| Google Cloud Generative AI Leader | Business level, no prerequisites | 90 minutes, 50–60 multiple choice, 99 USD plus tax, valid 3 years | AI adoption and governance roles |
| Microsoft AI-901: Azure AI Fundamentals | Fundamentals | Priced by country or region | Microsoft shops, first AI credential |
| Microsoft AI-103: Azure AI Apps and Agents Developer Associate | Associate | Proctored, may include interactive components | Engineers building AI apps in Python |
| Microsoft Applied Skills | Lab-based assessment | Varies by credential | Hands-on proof of a single skill |
How I judged them
A credential ranks higher here when it helps an IT pro do the job: governing AI tools, protecting data, automating workflows, supporting internal chat or search tools, and talking to security and engineering teams. It ranks lower when it’s mostly ML math, model training theory, or data science career prep.
1. AWS Certified AI Practitioner
AWS lists it as foundational, for people who want to validate foundational knowledge of AI/ML concepts and generative AI on AWS. The candidate role examples are broad: “cloud, development, data, IT, AI/ML, or line-of-business roles.” It’s 90 minutes, 65 questions, and 100 USD (AWS page).
It gives you enough AWS AI vocabulary for architecture, security, and procurement conversations, such as whether to use Amazon Bedrock or what IT should check before business teams connect data to an AI tool. It doesn’t prove you can build production AI apps.
Full review, including domain weights and recertification: Is AWS Certified AI Practitioner worth it for IT pros?
2. Google Cloud Generative AI Leader
Google says it’s “for anyone in any job role, with or without hands-on technical experience.” The exam covers gen AI fundamentals, Google Cloud’s gen AI offerings, techniques to improve model output, and business strategies for gen AI solutions. It runs 90 minutes, has 50–60 multiple choice questions, costs $99 plus tax, has no prerequisites, and is valid for 3 years (Google page).
It fits IT managers, Workspace admins supporting Gemini, and admins pulled into AI governance. It isn’t an implementation credential. Pair it with hands-on labs if you need practical exposure. Full review: Google Cloud Generative AI Leader for IT pros. For the wider Google path, see Google Cloud generative AI certification path.
3. Microsoft AI-901 (and why AI-900 and AI-102 are gone)
Microsoft retired AI-900 and AI-102 on June 30, 2026 (retired exams list). AI-901 is now the exam for Azure AI Fundamentals. Its study guide weights it 40–45% on AI concepts and 55–60% on implementing AI solutions with Microsoft Foundry, so it’s more hands-on than AI-900 was.
For Microsoft 365, Intune, and Entra admins, this is the natural first credential. Details: AI-900 retired: is AI-901 worth it for desktop engineers?
4. Microsoft AI-103, only if you build
AI-103 leads to Microsoft Certified: Azure AI Apps and Agents Developer Associate, and it validates designing, developing, and deploying Azure AI solutions “using Python and Microsoft Foundry” (Learn). The study guide weights generative AI and agentic solutions at 30–35%.
Caveat: Learn doesn’t explicitly call AI-103 the replacement for AI-102. It’s the active associate-level Azure AI developer path. If you don’t write Python or own an AI app, it’s too early. See AI-102 retired: is AI-103 worth it for Microsoft engineers?
5. Microsoft Applied Skills
Applied Skills credentials are lab-based assessments of one skill rather than a broad exam. They suit IT pros who want proof of a specific task before paying for a certification. Example: Build a generative AI chat app.
6. Coursera professional certificates
Coursera programs are structured courses, not vendor exams, so treat them as preparation rather than proof. Choose ones with labs, APIs, RAG, responsible AI, and cloud deployment content over academic model theory. Reviews of specific programs:
- Microsoft Generative AI Engineering certificate
- IBM RAG and Agentic AI professional certificate
- Google AI Professional Certificate
7. Specialty and vendor-neutral options
These fit narrower situations. Each one has its own review with checked facts:
- NVIDIA-Certified Associate: Generative AI LLMs. Associate level, 1 hour, 50–60 multiple choice questions, $125 (NVIDIA page). It’s relevant if your organization runs its own GPU inference. Review: NVIDIA Generative AI LLM Associate for IT pros.
- CompTIA AI Essentials. A non-cloud-specific option. Review: CompTIA AI Essentials for IT pros.
- IBM AI Engineering (Coursera). Project-based deep learning. Review: IBM AI Engineering professional certificate.
- AWS Certified Machine Learning - Specialty. Check its status on the AWS certification page before you plan around it; our review covers the retirement notice. Review: AWS ML Specialty for IT pros.
Putting a certification on your resume
List the outcome, not just the acronym. “AI-901 certified” says less than a line describing what you built or automated with what you learned. Use the full certification name and issuer on LinkedIn, and mention the credential in the project descriptions where it applied.
Recommended paths by role
- Desktop or endpoint engineer: AI-901, or an Applied Skills credential first. In an AWS-heavy environment, take AWS AI Practitioner.
- Sysadmin moving to cloud: follow your primary cloud. That’s AWS AI Practitioner, Generative AI Leader, or AI-901.
- IT manager or service desk lead: Generative AI Leader or AWS AI Practitioner, plus governance training for your stack (Microsoft 365 Copilot, Gemini, Bedrock).
- Automation-minded engineer: start with fundamentals and build a small internal tool. Take AI-103 once you know Azure is your platform.
FAQ
Is AI-900 still available?
No. It retired June 30, 2026. AI-901 is the current Azure AI Fundamentals exam.
Can I still take AI-102?
No. It retired June 30, 2026, along with the Azure AI Engineer Associate and its renewal assessment. AI-103 is the active associate-level Azure AI developer exam.
Which AI certification is easiest for IT pros?
The foundational ones: AWS AI Practitioner, Google Generative AI Leader, and AI-901. Choose based on your cloud, not difficulty.
Are Coursera certificates worth it?
As structured learning, yes, if they include labs. They don’t replace a vendor exam when an employer asks for one.
Sources
- AWS Certified AI Practitioner, AWS
- Generative AI Leader, Google Cloud
- Exam and assessment lab retirement, Microsoft Learn
- Study guide for Exam AI-901, Microsoft Learn
- Azure AI Apps and Agents Developer Associate, Microsoft Learn