I was studying for AZ-204. What now?

AZ-204 retired on 31 July 2026 and you can't book it. Most of the compute and messaging material carries over to AI-200. What's new is the AI data services work — vector storage and querying — and that's the heaviest domain, so re-weight your prep rather than restarting it.

Should I take AI-200 or AI-103?

AI-200 if you build and run the back end: containers, data, messaging, observability. AI-103 if you design the AI behaviour: models, agents, retrieval. Neither is a prerequisite for the other.

What score do I need?

700 or above on a 1 to 1,000 scale. It's scaled, so it isn't the same as getting 70% of questions right.

How many questions will I get?

Microsoft doesn't publish a fixed count. Its own guidance says most exams have between 40 and 60, that it varies, and that it changes. Anywhere you see an exact figure quoted, someone is guessing.

The exam assumes you write back-end code and Python is the assumed language. It isn't a syntax test, but the scenarios expect you to have built something.

Is there a prerequisite?

No formal one. The audience profile assumes real experience across the development lifecycle — requirements through deployment, security and monitoring.

How current are your questions?

Written against the skills outline dated 15 April 2026 and re-checked on a schedule. This exam launched in 2026 and the services it covers are still moving, so the review date is in the claims register.

Are these real exam questions?

No. Every item is original, written against Microsoft's published skills outline. Sites selling recalled questions put your certification at risk as well as their own.

What can I do without paying?

Untimed practice across all four domains, with every question explained. Timed mocks, a scaled score and your readiness range need a subscription.

Is Acedify affiliated with Microsoft?

No. We're independent and we don't administer this exam.

Microsoft · AI-200 · Azure AI Cloud Developer Associate

AZ-204 is retired. AI-200 is the back end of AI.

Microsoft retired AZ-204 on 31 July 2026 and replaced it with an exam about hosting AI in production: containers, vector-enabled data services, event-driven messaging, and the security and observability around them. Around two thirds of your AZ-204 knowledge still applies. The vector data work is the part that doesn't.

Questions weighted to the official domain split, so a mock matches the real mix
Every item tagged to the domain and subskill it tests
Readiness shown against 700, not as an invented percentage

Start practicing for free See what's included →

Free practice across all four domains. No card required.

AI-200Develop AI solutions by using Azure data management servicesHard

A catalogue API stores product documents in Azure Cosmos DB for NoSQL. Each document needs an embedding so the storefront can retrieve related items by meaning, and a downstream indexer must start when a new or updated item lands.

Which two actions should you take? Select two.

Two choices only. Your earliest pick was dropped.

0 of 2 selected

An original item written against the skills outline dated 15 April 2026. Not a real exam question.

AZ-204 to AI-200

Same developer. Different half of the stack.

AZ-204 tested whether you could build and ship applications on Azure. AI-200 tests whether you can host an AI application: package it into containers, store and query vectors, wire the messaging, and secure and observe the result. The compute and messaging ground is familiar. The AI data services are new to most Azure developers, and they carry the heaviest domain.

AZ-204 compared with AI-200
	AZ-204	AI-200
Status	Retired 31 Jul 2026	Current
Credential	Azure Developer Associate	Azure AI Cloud Developer Associate
Focus	Building applications on Azure	Hosting AI applications on Azure
New ground	—	Vector data services, AI-era observability

If you were part-way through AZ-204, most of it carries. Re-weight toward the data management domain rather than starting again. Microsoft retired the Azure Developer Associate renewal assessment with the exam, so there is no renewal path for the old credential.

Two new exams

AI-103 is the AI. AI-200 is what it runs on.

They arrived within weeks of each other and the names don't help. The split is cleaner than it looks: AI-103 tests the AI itself — model choice, agents, retrieval design, prompt and evaluation work on Microsoft Foundry. AI-200 tests the platform underneath — containers, vector storage, messaging, secrets and telemetry.

AI-103 compared with AI-200
	AI-103	AI-200
Replaces	AI-102	AZ-204
Tests	Building AI apps and agents	Hosting and operating AI apps
Heaviest domain	Generative AI and agentic solutions	Azure data management services
Take it if	You design the AI behaviour	You build and run the back end

If you came from AZ-204, AI-200 is the continuation of your path. If you came from AI-102, AI-103 is. Neither is a prerequisite for the other.

Scoring

700 is a pass. It isn't 70%.

Microsoft reports technical exams on a scale of 1 to 1,000, and 700 or above passes. It's a scaled score, so it doesn't correspond to answering 70% of questions correctly — how many you need depends on the difficulty of the questions you're given.

No penalty for a wrong answer, so never leave one blank. Some questions are worth more than one point, and on some exams you can't return to a question once you've moved past it.

How Microsoft scales a score1–1,000 · 700 to pass

You answered

72%of the questions correctly

700

Your scaled score lands somewhere around670–730Could go either way

Now change nothing except which form you were given

Same answers. Harder form, higher scaled score — that's what scaling is for. It's also why nobody can tell you the percentage you need.

Positions here are illustrative. Microsoft publishes the 1–1,000 scale and the 700 pass mark, but not the conversion from raw answers — which is the point. When a number can't be known exactly, a range is the honest way to report it.

The loop

Practice, sit, review, repeat

Step 1 is free. Steps 2 to 4 need a subscription.

certs.acedify.com/ai-200

Practice · untimed

Hard

Which two actions should you take? Select two.

Store embeddings on the items and run vector similarity search for semantic retrieval.

Turn on Azure Blob lifecycle management so unused embeddings expire after 30 days.

Set every vector query to strong consistency so nearest-neighbour recall improves.

Implement a change feed processor to detect new or updated items and kick off the indexer.

Embeddings plus vector search give you semantic retrieval in Cosmos DB. A change feed processor is how you react to new or updated items. Blob lifecycle and consistency level do not implement either. Domain: Develop AI solutions by using Azure data management services.

Click a step to see it.

What you get back

Not a prediction. A distance from the line.

Your readiness is a range, because it's an inference from the questions you've answered rather than a result you've earned. For a certification the outcome is binary, so the only question that matters is whether the whole range sits above 700.

While it straddles the line, you're a coin flip, and we'll say so.

AI-200 · 4 mocks completed

Where you stand

Borderline

530–720

Pass mark 700

Your range crosses the pass mark. On today's evidence it's a coin flip.

Three states

Try the three states. We won't show you a percentage chance of passing — nobody can calculate one honestly without years of real outcomes to calibrate against.

We won't show you a percentage chance of passing. Nobody can calculate one honestly without years of real outcomes behind it.

Domains

What AI-200 actually covers

Microsoft publishes a weighted skills outline. Our mocks follow those weights, so a domain worth a quarter of the exam gets a quarter of the questions.

AI-200 domain weights
Domain	Weight
Develop containerized solutions on Azure	20–25%
Develop AI solutions by using Azure data management services	25–30%
Connect to and consume Azure services	20–25%
Secure, monitor, and troubleshoot Azure solutions	20–25%

Weights are ranges because Microsoft publishes them as ranges. Skills measured as of 15 April 2026.

The weighting here is unusually flat — no domain is small enough to skip. On most exams you can afford one weak area. On this one, four domains between roughly a fifth and a third each means a gap anywhere costs you.

The data management domain carries the most marks and is the one least covered by prior Azure experience. If you came from AZ-204, that's where the work is: vector storage and querying rather than the compute and messaging you already know.

Read the claims register

Exam facts

Prices vary by country. Check Microsoft Learn for yours.

Exam: AI-200
Credential: Azure AI Cloud Developer Associate
Length: 120 minutes
Pass mark: 700 on a 1–1,000 scale
Delivery: Pearson VUE
Skills measured as of: 15 April 2026
Replaces: AZ-204, retired 31 July 2026

Read the claims register

Pricing

Practice free. Pay when you want to know where you stand.

Free

Untimed practice across all four domains
Every question explained, with its domain and subskill named

Pro

Timed mocks at exam length
A scaled score against the 700 pass mark
Domain-weighted readiness, and your range over time
Everything on every exam, one subscription

One subscription covers the Microsoft certifications, AWS, the Claude certifications, TOEFL, IELTS and the SAT.

The exam fee is separate and goes to Microsoft. We're prep, not registration.

Start practicing for free

FAQ

AI-200 questions

Find out which domain is keeping you under 700.

Free practice across all four domains, no card. The timed mocks and the readiness range are there when you want them.