Study Guide

AZ-204 is retired: What to study next for AI-200

Understand AZ-204 retirement, check your own credential expiry, and compare your existing preparation with the current AI-200 developer skills.

Updated September 20266 min readStudy GuideCloudCerty
Elizabeth Porter

Elizabeth Porter

CloudCerty Editorial Team

AZ-204 and the Azure Developer Associate renewal assessment have retired. Microsoft names Azure AI Cloud Developer Associate, earned through AI-200, as the replacement path. Use the current AI-200 skills to decide what to study next and retain older material where it still matches those tasks. Existing credentials remain active until their own expiry dates.

What changed after AZ-204 retired?

You can no longer schedule AZ-204, and the Azure Developer Associate credential and its renewal assessment are retired. Microsoft named the Azure AI Cloud Developer Associate (AI-200) as its replacement.

Microsoft’s AZ-204 study guide states that the exam retired on July 31, 2026, at 11:59 PM Central Standard Time. The credential page also states that the certification and renewal assessment are retired.

Microsoft’s July announcement names Azure AI Cloud Developer Associate (AI-200) as the replacement, available from July 9, 2026. Its current credential page and study guide define a separate exam and requirements; do not treat an AZ-204 result as automatic completion of AI-200.

An exam retirement does not, by itself, establish the retirement or current availability of every service in its old syllabus. Check current service documentation when reusing older learning material.

Sources used: Microsoft Certified: Azure Developer Associate - Certifications | Microsoft Learn — Credential overview and retirement warning; Study guide for Exam AZ-204: Developing Solutions for Microsoft Azure | Microsoft Learn — Retirement warning; skills measured as of January 14, 2026; Microsoft Certified: Azure AI Cloud Developer Associate - Certifications | Microsoft Learn — Current credential overview, exam duration, practice assessment status; Study guide for Exam AI-200: Developing AI Cloud Solutions on Azure | Microsoft Learn — Current skills measured and audience profile; July 2026 announcements - Partner Center announcements | Microsoft Learn — Retirement and replacement notice for AZ-204 and AI-200 availability

What happens to an existing credential?

An unexpired Azure Developer Associate credential remains active on your transcript until its own expiration date. Its renewal assessment is no longer available.

Microsoft’s current policy gives role-based certifications a one-year validity period after requirements are completed. An individual record may also reflect a renewal completed before retirement. Read the actual expiry date on your transcript rather than calculating it from the original award date.

Retirement does not cancel an unexpired credential early. The non-expiring Fundamentals policy does not apply to this Associate credential.

Sources used: Credential expiration policies | Microsoft Learn — Role-based certification expiry and retired credential transcript handling; Microsoft Certified: Azure Developer Associate - Certifications | Microsoft Learn — Credential overview and retirement warning; Microsoft Certification Renewal | Microsoft Learn — Six-month renewal window

Which AZ-204 topics can help with further study?

The archived English skills list dated January 14, 2026 covered compute, storage, security, monitoring, and service integration.

Its historical ranges were compute 25–30%, storage 15–20%, security 15–20%, monitoring and optimization 5–10%, and integration 20–25%. They describe the retired exam and must not be reused as AI-200 weights.

For existing notes, compare topics such as containers, App Service, Functions, Cosmos DB, Key Vault, managed identities, Service Bus, and Event Grid with the current AI-200 tasks. Keep explanations that still match, and check current documentation for changed capabilities or interfaces.

The historical candidate profile described programming and Azure tooling experience. That profile did not make AZ-900 a required prior credential, and it does not establish readiness for the current AI-focused exam.

Sources used: Microsoft Certified: Azure Developer Associate - Certifications | Microsoft Learn — Credential overview and retirement warning; Study guide for Exam AZ-204: Developing Solutions for Microsoft Azure | Microsoft Learn — Retirement warning; skills measured as of January 14, 2026; Study guide for Exam AI-200: Developing AI Cloud Solutions on Azure | Microsoft Learn — Current skills measured and audience profile

Managed identities: separate identity from permission

A managed identity can obtain Microsoft Entra tokens without the application managing those credentials. The target service must still authorize access.

A system-assigned managed identity belongs to one Azure resource and shares its lifecycle. It cannot be shared with other resources. A user-assigned identity is a separate Azure resource with an independent lifecycle and can be associated with multiple compute resources.

This distinction matters when a workload needs one identity to survive host replacement. It does not remove authorization requirements: the identity needs appropriate access to a target that supports Microsoft Entra authentication.

Sources used: Managed identities for Azure resources | Microsoft Learn — Identity types, lifecycle, sharing, and authorization; Study guide for Exam AZ-204: Developing Solutions for Microsoft Azure | Microsoft Learn — Retirement warning; skills measured as of January 14, 2026

Practice: choose an identity that survives host replacement

Use the stated requirements to compare the two identity types before reading the answer.

This original design exercise has complete fictional premises. It tests one identity distinction, not all Azure development skills or an actual exam question.

Try it yourself

A workload uses two Azure VM hosts that are periodically deleted and recreated. Both hosts must use the SAME stable managed identity to access an Azure Storage resource that supports Microsoft Entra authentication. The identity must survive host replacement, and existing permissions assigned to that identity should not need changing solely because a host is replaced. Does enabling a separate system-assigned identity on each host meet these requirements? Which type fits, and does creating it alone authorize storage access?

Show answer

Separate system-assigned identities do not meet the shared, independent-lifecycle requirement. A user-assigned managed identity fits. Creating it alone does not authorize access to storage.

Each system-assigned identity is tied to its parent resource and cannot be shared. A user-assigned identity can exist independently and be associated with both hosts, including replacement hosts. The workload must use that identity and have the appropriate approved permissions on the target. This choice does not remove other access requirements or allow authentication to arbitrary services that do not accept Microsoft Entra tokens.

Sources used: Managed identities for Azure resources | Microsoft Learn — Identity types, lifecycle, sharing, and authorization

What to study for the current AI-200 path

AI-200 focuses on AI back-end development on Azure. Compare its current profile and tasks with your background before choosing it.

The current profile includes Python, SDKs, data management, vector databases, containers, messaging, and monitoring. Existing Azure development knowledge can help, but it does not prove coverage of these skills.

The table summarizes the current outline. These are official percentage ranges, not exact question counts or required study-hour allocations. Read the full guide for its detailed tasks.

The current credential page gives 120 minutes for AI-200 and states that its practice assessment is not currently available. Check that page for updates and use its official learning and sandbox links. These details concern AI-200, not retired AZ-204.

Current AI-200 areaWeight rangeSuggested preparation
Containerized solutions20–25%Review ACR and tasks, App Service containers, Container Apps revisions and KEDA, and AKS. Explain the hosting and deployment choices in the listed tasks.
AI solutions using Azure data management services25–30%Review Cosmos DB for NoSQL vectors and change feed, PostgreSQL with pgvector, and Azure Managed Redis. Compare their documented roles against a stated data requirement.
Connect to and consume Azure services20–25%Review Service Bus, Event Grid, and Functions. Trace a small example’s event or message flow using the current learning material.
Secure, monitor, and troubleshoot solutions20–25%Review Key Vault, App Configuration, OpenTelemetry, and KQL. Explain identity and authorization separately and practice interpreting a small trace or query example.

Sources used: Microsoft Certified: Azure AI Cloud Developer Associate - Certifications | Microsoft Learn — Current credential overview, exam duration, practice assessment status; Study guide for Exam AI-200: Developing AI Cloud Solutions on Azure | Microsoft Learn — Current skills measured and audience profile; July 2026 announcements - Partner Center announcements | Microsoft Learn — Retirement and replacement notice for AZ-204 and AI-200 availability

Your next-study checklist

Use your intended role and the current AI-200 skills to decide what to keep and what to learn.

  • Check your actual transcript expiry date if you hold Azure Developer Associate.
  • Compare AI-200’s back-end development focus with the work you want to do.
  • Map each current task to material you understand or a gap needing study.
  • Review Python and AI data-management expectations when coming from older AZ-204 resources.
  • Work the managed-identity example and explain both lifecycle and authorization.
  • Use current official learning resources and check the credential page before booking AI-200.

Sources used: Credential expiration policies | Microsoft Learn — Role-based certification expiry and retired credential transcript handling; Microsoft Certified: Azure Developer Associate - Certifications | Microsoft Learn — Credential overview and retirement warning; Microsoft Certified: Azure AI Cloud Developer Associate - Certifications | Microsoft Learn — Current credential overview, exam duration, practice assessment status; Study guide for Exam AI-200: Developing AI Cloud Solutions on Azure | Microsoft Learn — Current skills measured and audience profile

Sources used

Fact-checked against Microsoft Learn, September 14, 2026:

Next steps

FAQ

Frequently Asked Questions

Practical answers to help you apply the guidance for Microsoft Certified Azure Developer Associate (AZ-204).

Can I reuse my AZ-204 notes for AI-200?
Use them where the concepts and tasks still match the current AI-200 guide. Check current documentation and add missing Python, vector-data, container, integration, and observability topics. A shared service name does not establish identical scope or an unchanged API.
Do I automatically receive the new credential because I passed AZ-204?
Do not assume conversion. Microsoft names AI-200 as the replacement path, and its current credential page sets its own exam requirement. An unexpired AZ-204 credential remains a separate record until its own expiry.

Keep Reading

Related Study Guides

Explore related guides and preparation topics.