AZ-204 Has Retired: The AI-200 Migration Guide for Azure Developers (2026)

Last updated: July 2026. Written and fact-checked by the Examinotion editorial team against Microsoft Learn certification pages, official skills outlines and Microsoft Tech Community announcements, all verified on 31 July 2026.

TL;DR Exam AZ-204 retired on 31 July 2026, and its replacement is AI-200: Developing AI Cloud Solutions on Azure. Roughly half the old syllabus survives. Identity, web apps, blob storage and API Management are gone, replaced by vector-capable data stores. Existing Azure Developer Associate credentials stay valid until their next expiry.

If you had AZ-204: Developing Solutions for Microsoft Azure on your 2026 plan, today is the day it stopped being available. Microsoft retired the exam, the Azure Developer Associate certification and its renewal assessments together, and pointed candidates at a new exam with a different name, a different shape and a noticeably different set of assumptions about what an Azure developer does. This guide sets out exactly what changed, what your existing credential is still worth, and how to redirect a part-finished AZ-204 study plan without wasting the work you have already done.

AZ-204 retired on 31 July 2026: what actually happened

Microsoft retired three things at once, not one. The warning on the Azure Developer Associate certification page reads: "This certification, related exam, and renewal assessments will retire on July 31, 2026. You will no longer be able to earn or renew this certification after this date." [1] The AZ-204 study guide adds the precise cut-off: "This exam will retire on July 31, 2026, at 11:59 PM Central Standard Time." [2]

The replacement is AI-200: Developing AI Cloud Solutions on Azure, which leads to the Microsoft Certified: Azure AI Cloud Developer Associate credential [4][5]. Microsoft's own retirement announcement pairs them directly, listing "Azure Developer Associate (AZ-204)" against replacement "AI-200" [6].

Two dates are worth noting because they caught people out. AZ-204 training retired in May 2026, two months ahead of the exam, so the instructor-led course was already gone while the exam was still bookable [6][20]. And there was no separate booking cut-off: the retirement date was the last day to sit, with the usual scheduling constraints applying, which are a maximum of two Microsoft exams booked at any one time and no booking more than 90 days ahead [11].

AZ-204 was not alone. Microsoft's retirement table also lists AZ-500 and PL-200 retiring on 31 August 2026, AZ-800 and AZ-801 on 30 September 2026, and MS-102 on 31 October 2026, with DP-100 having already gone on 1 June 2026 [3]. This is a broad refresh of the role-based catalogue, not a decision aimed at Azure developers specifically.

Does your existing Azure Developer Associate still count?

Yes, and this is the single most misunderstood point. Nothing you already hold was cancelled on 31 July 2026. Microsoft's policy language is explicit: "If you've already earned the certification or Applied Skills, it will stay on the transcript in your profile on Microsoft Learn." [3]

What stops is future activity. You cannot earn the certification any more, and you cannot renew it. Your credential runs to its own next expiry date and then lapses. Associate certifications expire annually, so if you renewed in March 2026 you hold a current Azure Developer Associate until March 2027 [13]. If you were inside your six-month renewal window and passed the renewal assessment before the deadline, you bought yourself one further year from the expiration date [13]. Microsoft recommended exactly that: "If you're eligible to renew your certification before the retirement date, we recommend that you consider doing so; you will not have an option to do so after the certification retires." [3]

Microsoft Partners get a separate concession. Those who earned an eligible certification before it retires "will continue to earn points or credit toward Partner requirements for one year after the retirement date." [3]

For the full mechanics of how Microsoft renewals work, including the six-month window and the free unproctored assessment, see our Microsoft AI certification renewal guide.

The AZ-400 prerequisite gap nobody is talking about

Here is a genuine, checkable gap that most coverage has missed. The Microsoft Certified: DevOps Engineer Expert certification requires one of two associate credentials as a prerequisite. As of that page's last update on 27 July 2026, those two are still, verbatim, "Microsoft Certified: Azure Administrator Associate, Microsoft Certified: Azure Developer Associate certification." [10]

AI-200 and the Azure AI Cloud Developer Associate are not yet listed on that page [10]. If you passed AZ-204 before it retired, your DevOps Engineer Expert route is intact. If you are starting from scratch today, the only documented path into AZ-400 is currently AZ-104 and the Azure Administrator Associate.

Treat this as a timing gap rather than a permanent exclusion. It is normal for prerequisite lists to lag a new exam by a few weeks, and there is every reason to expect AI-200 to be added. But if AZ-400 is on your roadmap this quarter, plan around what the certification page actually says today rather than what it is likely to say in September.

What AI-200 covers

AI-200: Developing AI Cloud Solutions on Azure is an intermediate, developer-role exam [5]. Microsoft describes the audience as follows: "As a candidate for this Microsoft Certification, you're responsible for contributing to all phases of implementing AI solutions on Azure, with an emphasis on back-end services and components." [4]

The proficiency list is where the change of emphasis shows. Microsoft says you should be proficient in Azure SDKs and third-party SDKs, Azure data management services, Azure monitoring and troubleshooting, Azure messaging and eventing, vector databases, Python programming, and implementing containerised applications on Azure [4][5].

Compare that with what AZ-204 assumed and you can see the repositioning. AZ-204 asked for around two years of programming experience in a language of your choice, plus proficiency with Azure CLI and Azure PowerShell, app authentication and authorisation, and data storage options [2]. AI-200 names Python specifically and adds vector databases as a baseline expectation, while dropping the identity and CLI assumptions entirely [4].

Sources for this table: [1][2][4][5][7][12].

The four AI-200 domains and their weightings

The skills outline splits into four domains [4]:

Develop containerized solutions on Azure covers Azure Container Registry including image build, storage, versioning and management plus ACR Tasks, deploying containers to Azure App Service with environment variables and secrets, Azure Container Apps including environment configuration and revision management, event-driven scaling with Kubernetes Event-driven Autoscaling (KEDA), and deploying and managing applications on Azure Kubernetes Service (AKS) using manifest files [4].

Develop AI solutions by using Azure data management services is the heaviest domain and the most changed. It covers Azure Cosmos DB for NoSQL (SDK queries, Request Unit optimisation through indexing policies and consistency levels, storing and retrieving embeddings, vector similarity search, and change feed processors), Azure Database for PostgreSQL (schema modelling, indexing strategies, reducing pgvector compute overhead, sizing compute and memory for vector workloads, and retrieval-augmented generation patterns with metadata filters), and Azure Managed Redis (caching, expiration, invalidation and vector indexing) [4].

Connect to and consume Azure services covers Azure Service Bus including dead-letter queue handling, topics and subscriptions, Azure Event Grid including filters, custom events and retries, and Azure Functions for serverless APIs with triggers and bindings [4].

Secure, monitor, and troubleshoot Azure solutions covers Azure Key Vault secret rotation and retrieval, Azure App Configuration, distributed tracing with OpenTelemetry SDKs, and writing KQL queries to analyse logs and metrics [4].

One structural note that matters for study planning: the AI-200 study guide has no "Skills measured as of" date and no change log [4], unlike AZ-204 which carried "Skills measured as of January 14, 2026" [2]. That is normal for a first-generation exam, but it means there is no prior version to calibrate against.

What carried over from AZ-204, and what quietly disappeared

This is the section to read if you are part-way through an AZ-204 study plan, because it tells you which revision to keep and which to abandon.

What survived

What is gone entirely

These AZ-204 topics have no counterpart anywhere in the AI-200 skills outline [2][4]:

1. Azure App Service Web Apps as a domain. Creating web apps, diagnostics and logging, TLS and API settings, autoscaling and deployment slots are all gone. Only container deployment to App Service survives.
2. Azure Blob Storage, in full. Properties and metadata, SDK data operations, storage policies and data lifecycle management.
3. Azure Queue Storage.
4. Azure Container Instances. Replaced in scope by AKS and Container Apps.
5. Azure API Management, in full. This was an entire sub-domain covering instance creation, API documentation, access configuration and policies.
6. Azure Event Hubs. Event Grid survives, Event Hubs does not.
7. The whole identity and authorisation sub-domain. Microsoft identity platform, Microsoft Entra ID user and app authentication, shared access signatures and Microsoft Graph integration.
8. Application Insights instrumentation, availability tests and alerts as named objectives. Instrumentation now appears as OpenTelemetry instead.

Managed identities deserve a careful note. They appear in AI-200's training materials but not in the exam's skills outline [4][15]. That makes them no longer a named exam objective, which is not the same as saying managed identity has stopped mattering in real Azure development.

What is genuinely new

The new material clusters almost entirely around vector-capable data stores. Azure Database for PostgreSQL with pgvector is the single largest new block at six separate bullet points, covering schema design, indexing strategy, pgvector compute overhead, resource sizing for vector workloads, RAG patterns with metadata filters, and connection optimisation [4]. Nothing in AZ-204 resembles it.

Azure Managed Redis is new, and note the product name carefully, because AZ-204's materials referred to Azure Cache for Redis while AI-200's objectives say Azure Managed Redis [2][4]. Azure Kubernetes Service is new to this exam family, as are ACR Tasks, KEDA, OpenTelemetry SDKs and explicit Request Unit optimisation [4].

The honest one-line summary is that AI-200 is narrower and deeper than AZ-204, not "AZ-204 with AI bolted on". Microsoft dropped identity, web apps, blob storage and API gateways, kept the compute, messaging and secrets spine, and spent the reclaimed weight on vector data. Observability also got a real promotion: it was worth 5-10% in AZ-204 and now shares a 20-25% domain with security [2][4].

A trap inside Microsoft's own study guide

Be careful with one thing. The AI-200 study guide's "Study resources" table is stale AZ-204 boilerplate that Microsoft has not yet updated. Its "Get trained" link points at the AZ-204 exam page, its "Find documentation" row lists Blob Storage, API Management, Event Hubs, Queue Storage and App Service, and its "Find a video" row filters the Exam Readiness Zone by AZ-204 [4].

None of those services appear in the AI-200 skills measured. Only the "Skills measured" section of that page is authoritative. Anyone revising from the resources table will study four topics that cannot appear on the exam.

AI-200 versus AI-103: two different Azure AI developer roles

Microsoft now runs two intermediate Azure AI developer exams, and candidates routinely confuse them. They are complementary, not alternative.

Sources: [3][4][5][6][8][9].

The positioning is clearest in Microsoft's own descriptions. AI-200 "validates your ability to design, build, and implement AI solutions on Azure, with a focus on back-end services, scalable architectures, and the full development lifecycle" [5]. AI-103 "validates your expertise in designing, developing, and deploying advanced Azure AI solutions using Python and Microsoft Foundry" [9].

The simplest framing is that AI-200 is the plumbing and AI-103 is the model layer. AI-200 examines the containers, queues, vector databases, secrets and traces that an AI application runs on, and never lists a model, a prompt, an agent or responsible AI as an objective. AI-103 examines model selection, agentic solutions, computer vision, text analysis and information extraction, and never lists AKS, Service Bus, Key Vault or OpenTelemetry [4][8]. There is essentially zero objective-level overlap.

One shared term to watch. Both outlines use RAG, but they mean different work. In AI-200 it means implementing retrieval patterns inside PostgreSQL and Cosmos DB. In AI-103 it means designing the ingestion, grounding and evaluation pipeline [4][8].

Note the naming currency too. Azure AI Foundry is now Microsoft Foundry, and the AI-103 certification page tags the product accordingly [9][18]. Study material still calling it Azure AI Foundry is out of date.

If AI-103 is the exam you actually need, Examinotion sells practice tests for it. Start with our AI-103 exam preparation and the AI-103 study guide, or read how to pass the AI-103 exam and our guide to the AI-102 retirement and AI-103 successor path.

Which exam should you sit now?

A short decision framework, based on the objectives rather than the marketing.

If you are earlier in your journey than any of these rows assume, our role-based decision guide to Microsoft AI certifications and the Microsoft AI certification roadmap are the better starting points. For the fundamentals-versus-associate question specifically, see AI-901 versus AI-103.

Is AI-200 harder than AZ-204? An honest read

There is no credible candidate-sentiment data yet, and any page claiming a pass rate for AI-200 is guessing. The exam has been generally available for weeks rather than months, there is no official practice assessment to calibrate against, and no body of first-hand reports from identifiable practitioners exists. We would rather say that plainly than invent a number.

What we can do is read Microsoft's own published structure, and it is genuinely informative.

Microsoft publishes a duration bracket table [12]:

AZ-204 was 100 minutes. AI-200 is 120 minutes [1][5]. By Microsoft's own bracketing, AZ-204 sat in the "without labs" tier and AI-200 sits in the "may contain labs" tier. Microsoft will not confirm labs for any specific exam, stating: "Because labs can be removed at any time due to Azure outages, bandwidth issues, etc., Microsoft does not provide a list of exams with labs." [12] The certification page separately confirms AI-200 "may have interactive components to complete as part of this exam" [5].

So the accurate formulation is this. AI-200 is scheduled at the duration Microsoft reserves for associate exams that may contain labs, twenty minutes longer than AZ-204, and Microsoft confirms interactive components are possible without committing to labs in advance.

Three further signals point the same way:

• Python is assumed, not taught. For a C# and .NET developer coming from AZ-204, that is new ground, and Microsoft states it as a proficiency expectation rather than an exam objective [4].
• English only, against AZ-204's ten languages [1][5]. For non-native speakers this is a material step up, partly offset by the accommodation that lets you request an additional 30 minutes when an exam is not offered in your preferred language [4].
• No practice assessment and no change log, so there is no official calibration tool and no previous version to benchmark against [4][7].

The one substantive Microsoft comment on scope came in the reply thread under the retirement announcement, where a contributor argued that retiring AZ-204 leaves a gap for .NET shops building microservice back-ends on App Service and Blob Storage. A Microsoft respondent replied that the new course "covers similar ground as AZ-204, but with a greater focus on code and observability" [17]. The domain-by-domain comparison above broadly supports that framing, with the important caveat that the vector-data block is genuinely new rather than a reframing of anything.

For how the 700 of 1000 threshold actually works, and why it is not 70 percent, see our explainer on the Microsoft exam passing score.

AI-200 exam format, booking and scoring

The essentials, all from Microsoft's own pages [4][5][11][12]:

• Duration: 120 minutes, 140 minutes of seat time
• Passing score: 700 or greater, out of 1000
• Question count: not published. Microsoft's general guidance is that most exams "typically contain between 40-60 questions" [12]
• Languages: English only
• Delivery: proctored, with interactive components possible
• Prerequisites: none
• Retakes: 24 hours after a first attempt, longer for subsequent attempts
• Price: Microsoft publishes no figure, stating only "Price based on the country or region in which the exam is proctored" [5]. Check Pearson VUE at the point of booking rather than trusting a number quoted on a blog

Booking runs through your Microsoft Learn profile and then Pearson VUE [5][11]. One warning from Microsoft is worth repeating because it costs people their entire transcript: "We strongly recommend that you register for an exam with a personal MSA account. If you register with an organizational (work/school) AAD account, your exam records will be lost if you leave your organization and they will be unrecoverable." [5]

If you book and then need to move the date, the 24-hour rule governs: "Be sure to reschedule or cancel exam appointments at least 24 hours before your scheduled exam or your fee will be forfeited. If you used a company-purchased voucher, the voucher will be forfeited." [16]

AI-200 is an associate role-based exam, which means you can access Microsoft Learn during the exam, covering everything on learn.microsoft.com except Q&A, practice assessments and your own profile. No extra time is added and the clock keeps running [12]. Microsoft's own caution is worth taking seriously: "It is not something you should be leveraging to answer every question. If you do, you won't be able to complete the exam in the time allotted, and that is by design." [12] Our guide to using Microsoft Learn during your exam covers how to use that access without burning your time budget, and the online proctored exam guide covers the OnVUE check-in process.

Preparing while no official practice assessment exists

This is the awkward part of sitting a first-generation exam, and it is worth stating precisely. Microsoft's certification page says: "The Practice Assessment for this exam is not currently available. Practice Assessments are usually available within 8 weeks of the exam being out of beta and generally available." [5]

That is corroborated by Microsoft's master list of available practice assessments, which was updated on 30 July 2026, runs to 49 entries, and contains neither AI-200 nor AI-103 [7]. AZ-204 is still listed there today and will disappear along with the exam [7].

What does exist right now is substantial and free:

• Course AI-200T00-A: Develop AI cloud solutions on Azure, a five-day instructor-led course tagged directly to the certification, available in twelve languages [15].
• Nine free self-paced learning paths totalling 24 modules on Microsoft Learn, which are the same paths the instructor-led course is built from [15][21]. They map cleanly onto the four domains: three paths for containerised solutions, three for the data management services domain, one for connecting to and consuming Azure services, and two for security, monitoring and troubleshooting.
• The exam sandbox at aka.ms/examdemo, linked from the certification page, which demonstrates the exam interface and question types [4][5]. It contains no AI-200 content, so treat it as a familiarisation tool rather than practice.

One readiness check before you commit. Several of those learning paths state "Programming experience with Python" as a flat prerequisite, and the Cosmos DB and PostgreSQL paths additionally require an "Understanding of machine learning concepts including embeddings and similarity search" [21]. If neither is true of you yet, build that foundation before booking rather than discovering the gap in the exam.

A study plan for part-finished AZ-204 candidates

If you were mid-preparation when the retirement landed, here is how to salvage the work.

1. Keep everything you revised on Container Registry, Container Apps, Functions, Service Bus, Event Grid, Key Vault, App Configuration, Cosmos DB and KQL. That material transfers almost intact, and in several cases AI-200 goes deeper than AZ-204 did.
2. Stop revising App Service Web Apps beyond container deployment, Blob Storage, Queue Storage, Container Instances, API Management, Event Hubs and the entire identity and authorisation block. None of it is examinable on AI-200 [2][4].
3. Add the vector data block first, because at 25-30% it is both the heaviest domain and the one where you have no prior credit. Work through the Cosmos DB, PostgreSQL and Managed Redis learning paths in that order.
4. Add AKS, ACR Tasks and KEDA next. If your container experience stopped at Container Apps, this is your second largest gap.
5. Convert your Application Insights knowledge to OpenTelemetry, which is now the named skill.
6. Verify your Python, honestly. If your Azure work has been C# throughout, treat this as a real workstream rather than a footnote.

For a worked example of how we structure a month of preparation, our AI-103 30-day study plan uses the same week-by-week approach and is easy to adapt to AI-200's four domains.

Mistakes AZ-204 veterans make when moving to AI-200

• Revising from the study guide's resources table. It still lists AZ-204 services that cannot appear on AI-200 [4].
• Assuming containers means the same containers. Azure Container Instances was examinable on AZ-204 and is absent from AI-200, which uses ACR, App Service containers, Container Apps and AKS instead [2][4].
• Carrying over identity revision. Entra ID authentication, shared access signatures and Microsoft Graph were a substantial AZ-204 block and are simply not on AI-200 [2][4].
• Saying "Azure Cache for Redis" when the objectives say Azure Managed Redis [4]. Product naming currency matters on Microsoft exams.
• Treating vector search as background reading. It is the heaviest domain, spread across three separate data services [4].
• Assuming AI-200 unlocks AZ-400. Not yet, on the evidence of the certification page as at 27 July 2026 [10].
• Budgeting AZ-204's 100 minutes. AI-200 runs to 120 with interactive components possible [1][5][12].

Where AI-200 sits in the wider Microsoft AI certification path

AI-200 is the platform-side associate exam in a family that Microsoft rebuilt substantially through the first half of 2026. AI-102 and the Azure AI Engineer Associate retired on 30 June 2026 and were replaced by AI-103 [3][6]. AI-900 retired on the same date and was replaced by AI-901 [3][6]. Now AZ-204 has been replaced by AI-200.

Read together, the pattern is clear enough: Microsoft has folded AI workload assumptions into the developer and engineer tracks rather than keeping AI as a separate specialism bolted onto a general Azure certification. If you are mapping a multi-year path rather than a single exam, our Microsoft AI certification roadmap sets out how the fundamentals, associate and expert tiers now connect, and is AI-103 worth it covers the career case for the adjacent developer credential.

One point of transparency. Examinotion does not currently sell AI-200 preparation. Our practice tests cover AB-730, AB-731, AB-900, AB-100, AI-901 and AI-103, and you can compare them on the Microsoft exams hub. If AI-103 is the exam that actually fits your role, that is where we can help today.

Frequently asked questions

When did AZ-204 retire?

AZ-204 retired on 31 July 2026 at 11:59 PM Central Standard Time. Microsoft retired the exam, the Azure Developer Associate certification and its renewal assessments together on that date. Related instructor-led training had already been withdrawn in May 2026, two months ahead of the exam itself [2][3][6].

What replaced AZ-204?

AI-200: Developing AI Cloud Solutions on Azure replaced AZ-204. Passing it earns the Microsoft Certified: Azure AI Cloud Developer Associate credential. It is an intermediate developer-role exam of 120 minutes, generally available since July 2026, and it carries no prerequisites [4][5][6].

Is my AZ-204 certification still valid after retirement?

Yes. Any Azure Developer Associate certification you already earned stays on your Microsoft Learn transcript and remains current until its own expiry date. What ends is the ability to earn it or renew it. Once it expires after 31 July 2026, there is no renewal route available [1][3].

Is AI-200 harder than AZ-204?

There is no credible pass-rate data yet. Structurally, AI-200 runs 120 minutes against AZ-204's 100, which places it in Microsoft's bracket for associate exams that may contain labs. It also assumes Python specifically, is offered in English only, and has no practice assessment [1][5][12].

What is the difference between AI-200 and AI-103?

AI-200 covers the platform an AI application runs on: containers, messaging, vector databases, secrets and observability. AI-103 covers the model layer: generative AI, agents, computer vision and text analysis in Microsoft Foundry. Their skills outlines have essentially no objective-level overlap, so they are complementary rather than alternatives [4][8].

Does AI-200 count as a prerequisite for AZ-400?

Not as of 27 July 2026. The DevOps Engineer Expert certification page still lists only Azure Administrator Associate and Azure Developer Associate as qualifying prerequisites. Expect AI-200 to be added, but plan around the published list rather than the expectation if your timeline is short [10].

Is there a practice assessment for AI-200?

No. Microsoft's certification page states the practice assessment "is not currently available" and that these usually arrive within eight weeks of an exam becoming generally available. Microsoft's master list of practice assessments, updated 30 July 2026, confirms AI-200 is absent [5][7].

Do I need Python for AI-200?

Effectively yes. Microsoft lists Python programming among the proficiencies you should already have, and several of the official learning paths state Python experience as a flat prerequisite. It is treated as an assumption rather than something the exam teaches, so a .NET-only background needs genuine ramp-up time [4][21].

Where to go from here

AZ-204's retirement is a redirection, not a loss. Around half of the syllabus you were revising still counts, the credential you already hold keeps its value until it expires, and the new material concentrates in one identifiable place: vector-capable data stores. Work the free learning paths, verify your Python honestly, and book once the vector domain feels solid rather than rushing a first-generation exam with no practice assessment behind it.

If the model layer rather than the platform is where your work actually sits, AI-103 is the exam to target, and Examinotion's practice tests cover it in full. Browse Examinotion's exam preparation courses on the Microsoft exams hub, or go straight to AI-103 practice tests to see where you stand.

Sources

1. Microsoft Certified: Azure Developer Associate - Microsoft Learn, page updated 16 March 2026, accessed 2026-07-31
2. Study guide for Exam AZ-204: Developing Solutions for Microsoft Azure - Microsoft Learn, skills measured as of 14 January 2026, accessed 2026-07-31
3. Exam and assessment lab retirement - Microsoft Learn, page updated 27 July 2026, accessed 2026-07-31
4. Study guide for Exam AI-200: Developing AI Cloud Solutions on Azure - Microsoft Learn, page updated 5 May 2026, accessed 2026-07-31
5. Microsoft Certified: Azure AI Cloud Developer Associate - Microsoft Learn, page updated 9 July 2026, accessed 2026-07-31
6. The AI job boom is here. Are you ready to showcase your skills? - Mark Winters, Microsoft Tech Community, published 3 March 2026, accessed 2026-07-31
7. Practice Assessments for Microsoft Certifications - Microsoft Learn, page updated 30 July 2026, accessed 2026-07-31
8. Study guide for Exam AI-103: Developing AI Apps and Agents on Azure - Microsoft Learn, skills measured as of 16 April 2026, accessed 2026-07-31
9. Microsoft Certified: Azure AI Apps and Agents Developer Associate - Microsoft Learn, page updated 23 July 2026, accessed 2026-07-31
10. Microsoft Certified: DevOps Engineer Expert - Microsoft Learn, page updated 27 July 2026, accessed 2026-07-31
11. Register and schedule an exam - Microsoft Learn, page updated 2 March 2026, accessed 2026-07-31
12. Exam duration and exam experience - Microsoft Learn, page updated 4 June 2026, accessed 2026-07-31
13. Microsoft Certification Renewal - Microsoft Learn, accessed 2026-07-31
14. Renewal for Microsoft Certified: Azure Developer Associate - Microsoft Learn, page updated 16 March 2026, accessed 2026-07-31
15. Course AI-200T00-A: Develop AI cloud solutions on Azure - Microsoft Learn, page updated 6 July 2026, accessed 2026-07-31
16. Reschedule and cancellation policies - Microsoft Learn, accessed 2026-07-31
17. Reply thread on the certification retirement announcement - Microsoft Tech Community, replies dated 5 to 8 March 2026, accessed 2026-07-31
18. What is Microsoft Foundry? - Microsoft Learn, accessed 2026-07-31
19. New Microsoft Certified: Azure AI Cloud Developer Associate Certification - LibertyMunson, Microsoft Tech Community, published 5 May 2026, accessed 2026-07-31
20. Microsoft Exam Updates: Certification Changes and Policies - Pearson VUE, accessed 2026-07-31
21. Learning paths for Course AI-200T00-A - Microsoft Learn, nine paths and 24 modules, path pages updated between November 2025 and February 2026, accessed 2026-07-31