Exam AI-103: Developing AI Apps and Agents on Azure. This page shows what changed in the study guide, which Microsoft's own page does not. Microsoft's page only ever shows one version, so a version that is still in force can disappear from it before a new one takes effect.
AI-103 study guide on Microsoft LearnThe official guide: skills measured, audience profile and exam details.AI-103 Study MaterialsCourses, videos, practice tests and more for this exam.Changes
The Current tab shows the guide as it is today, and History will list every change we record.
Compared to AI-102
AI-103 measures 64 skills; AI-102 measures 106. We matched every skill of both guides by hand on November 16, 2026. Compared with AI-102, 0 skills are the same and 18 are reworded (all 18 now in another group or domain); 29 skills of AI-103 replace 58 of AI-102 that were merged or split; 17 are new; and 30 skills of AI-102 are not in AI-103. Most of the guide is different, so start with the domains below.
What is different at a glance
Domains
How the areas changed
Judged by reading the two guides side by side. Click an area to jump to it in the full guide below.
Show every difference
Green underlined text is new or reworded. Red strikethrough shows old wording that was removed or replaced, not necessarily a topic that disappeared. Unmarked text in a one-to-one comparison is unchanged. Every skill of both guides was matched by hand, by what it asks of a candidate and not only by its words. A grey note after a skill says where it moved to or came from. A merged or split skill is shown in its current wording with the previous skills struck through below it; read the note for retained topics, new requirements and details no longer named. This is the whole AI-103 study guide, so you can see each difference in place.
AI-103 skills measured, compared to AI-102
Audience profile
As a candidate for this Microsoft Certification, you’re an Azure AI engineer who builds, manages, and deploys agents and AI solutions that take advantage of Microsoft Foundry.
For this exam, you should have experience developing apps by using Python, and you need to be familiar with the capabilities of general AI, generative AI, and Azure services.
Your responsibilities include:
- Planning and managing Azure AI solutions.
- Implementing generative AI and agentic solutions.
- Implementing computer vision solutions.
- Implementing text analysis solutions.
- Implementing information extraction solutions.
In this role, you collaborate with business stakeholders, solution architects, data scientists, DevOps engineers, and cloud security engineers to design, implement, and maintain AI solutions.
As a Microsoft Azure AI engineer, you build, manage, and deploy AI solutions that leverage Azure AI.
Your responsibilities include participating in all phases of AI solutions development, including:
Requirements definition and designDevelopmentDeploymentIntegrationMaintenancePerformance tuningMonitoring
You work with solution architects to translate their vision. You also work with data scientists, data engineers, Internet of Things (IoT) specialists, infrastructure administrators, and other software developers to:
Build complete and secure end-to-end AI solutions.Integrate AI capabilities in other applications and solutions.
As an Azure AI engineer, you have experience developing solutions that use languages such as:
PythonC#
You should be able to use Representational State Transfer (REST) APIs and SDKs to build secure image processing, video processing, natural language processing, knowledge mining, and generative AI solutions on Azure. You should:
Understand the components that make up the Azure AI portfolio and the available data storage options.Be able to apply responsible AI principles.
Skills at a glance
- Plan and manage an Azure AI solution (
20–25%25–30%) - Implement generative AI and agentic solutions (30–35%) — merged from Implement generative AI solutions and Implement an agentic solution
- Implement computer vision solutions (10–15%)
- Implement
natural language processingtext analysis solutions (15–20%10–15%) - Implement
knowledge mining andinformation extraction solutions (15–20%10–15%) Implement generative AI solutions (15–20%)— merged into Implement generative AI and agentic solutionsImplement an agentic solution (5–10%)— merged into Implement generative AI and agentic solutions
Plan and manage an Azure AI solution (20–25% 25–30%)
Choose the appropriate Foundry services for generative AI and agents
- Choose an appropriate model for each task, including large language models (LLMs), small language models, multimodal models, and Foundry Tools ← merged from 5 skills of AI-102:
Select the appropriate service for a computer vision solution(from Select the appropriate Microsoft Foundry Services)Select the appropriate service for a natural language processing solution(from Select the appropriate Microsoft Foundry Services)Select the appropriate service for a speech solution(from Select the appropriate Microsoft Foundry Services)Select the appropriate service for an information extraction solution(from Select the appropriate Microsoft Foundry Services)Choose the appropriate AI models for your solution(from Plan, create and deploy a Microsoft Foundry Service)
SelectChoose the appropriateserviceFoundry services foragenerativeAI solutiontasks, grounding, vector search, agent workflows, or multimodal processing ← from Select the appropriate Microsoft Foundry ServicesGenerative service selection now also covers grounding, vector search, agent workflows and multimodal processing.Select theChoose an appropriateservicemethod fora knowledge mining solutionretrieval and indexing ← from Select the appropriate Microsoft Foundry ServicesKnowledge mining becomes choosing a retrieval and indexing method.- Choose appropriate memory, tool, and knowledge integration services for agent solutions
Set up AI solutions in Foundry
- Design Azure infrastructure for AI apps and agent-based solutions ← merged from 5 skills of AI-102:
Create an Azure AI resource(from Plan, create and deploy a Microsoft Foundry Service)Plan and prepare for a generative AI solution(from Implement generative AI solutions)Deploy a hub, project, and necessary resources with Microsoft Foundry(from Implement generative AI solutions)Provision an Azure OpenAI in Foundry Models resource(from Implement generative AI solutions)Configure the necessary resources to build an agent(from Implement an agentic solution)
- Choose appropriate deployment options ← one of 2 skills split from 1 skill of AI-102:
Deploy AI models using the appropriate deployment options(from Plan, create and deploy a Microsoft Foundry Service)
- Configure model and agent deployments ← part of the same change as the skill above
- Integrate
MicrosoftFoundryServices into aprojects with continuous integration and continuousdeliverydeployment (CI/CD)pipelinepipelines ← from Plan, create and deploy a Microsoft Foundry Service
Select the appropriate Microsoft Foundry Services
Select the appropriate service for a generative AI solution→ now in Choose the appropriate Foundry services for generative AI and agentsSelect the appropriate service for a computer vision solution→ merged with 4 other skills into one skill in Choose the appropriate Foundry services for generative AI and agentsSelect the appropriate service for a natural language processing solution→ merged with 4 other skills into one skill in Choose the appropriate Foundry services for generative AI and agentsSelect the appropriate service for a speech solution→ merged with 4 other skills into one skill in Choose the appropriate Foundry services for generative AI and agentsSelect the appropriate service for an information extraction solution→ merged with 4 other skills into one skill in Choose the appropriate Foundry services for generative AI and agentsSelect the appropriate service for a knowledge mining solution→ now in Choose the appropriate Foundry services for generative AI and agents
Plan, create and deploy a Microsoft Foundry Service
Plan for a solution that meets Responsible AI principlesCreate an Azure AI resource→ merged with 4 other skills into one skill in Set up AI solutions in FoundryChoose the appropriate AI models for your solution→ merged with 4 other skills into one skill in Choose the appropriate Foundry services for generative AI and agentsDeploy AI models using the appropriate deployment options→ split into 2 skills in Set up AI solutions in FoundryInstall and utilize the appropriate SDKs and APIs→ regrouped with 3 other skills into 2 skills in Implement generative AI and agentic solutionsDetermine a default endpoint for a service→ regrouped with 3 other skills into 2 skills in Implement generative AI and agentic solutionsIntegrate Microsoft Foundry Services into a continuous integration and continuous delivery (CI/CD) pipeline→ now in Set up AI solutions in FoundryPlan and implement a container deployment
Manage, monitor, and secure a Microsoft Foundry Service AI systems
- Manage quotas, scaling, rate limits, and cost footprints for model and agent workloads ← merged from 2 skills of AI-102:
Manage costs for Microsoft Foundry ServicesOptimize and manage resources for deployment, including scalability and foundational model updates(from Implement generative AI solutions)
- Monitor model performance, drift, safety events, and grounding quality ← one of 2 skills that replace 3 skills of AI-102:
Monitor an Azure AI resourceConfigure model monitoring and diagnostic settings, including performance and resource consumption(from Implement generative AI solutions)Enable tracing and collect feedback(from Implement generative AI solutions)
- Monitor data ingestion quality, search index health, and relevance performance
- Configure security, including managed identity, private networking, keyless credentials, and role policies ← merged from 2 skills of AI-102:
Manage and protect account keysManage authentication for a Microsoft Foundry Service resource
Implement responsible AI across generative AI and agentic systems
- Configure safety filters, guardrails, risk detection, and content moderation ← merged from 3 skills of AI-102:
Implement content moderation solutions(from Implement AI solutions responsibly)Implement responsible AI, including content filters and blocklists(from Implement AI solutions responsibly)Prevent harmful behavior, including prompt shields and harm detection(from Implement AI solutions responsibly)
ConfigureApply responsible AIinsights,instrumentation, includingcontentevaluators, safety evaluations, and explanation tooling ← from Implement AI solutions responsiblyResponsible AI insights become instrumentation with evaluators, safety evaluations and explanation tooling.- Implement auditing through trace logging, provenance metadata, and approval workflows ← one of 2 skills split from 1 skill of AI-102:
Design a responsible AI governance framework(from Implement AI solutions responsibly)
- Govern agent behavior with oversight modes, constraints, and tool-access controls ← part of the same change as the skill above
Implement AI solutions responsibly
Implement content moderation solutions→ merged with 2 other skills into one skill in Implement responsible AI across generative AI and agentic systemsConfigure responsible AI insights, including content safety→ now in Implement responsible AI across generative AI and agentic systemsImplement responsible AI, including content filters and blocklists→ merged with 2 other skills into one skill in Implement responsible AI across generative AI and agentic systemsPrevent harmful behavior, including prompt shields and harm detection→ merged with 2 other skills into one skill in Implement responsible AI across generative AI and agentic systemsDesign a responsible AI governance framework→ split into 2 skills in Implement responsible AI across generative AI and agentic systems
Implement generative AI and agentic solutions (30–35%)
Build generative applications by using Foundry
- Deploy and consume LLMs, small models, code models, and multimodal models ← one of 2 skills that replace 4 skills of AI-102:
Deploy the appropriate generative AI model for your use case(from Implement generative AI solutions)Select and deploy an Azure OpenAI model(from Implement generative AI solutions)Submit prompts to generate code and natural language responses(from Implement generative AI solutions)Use large multimodal models in Azure OpenAI(from Implement generative AI solutions)
- Implement
a RAG pattern by grounding a modelretrieval-augmented generation (RAG) inyour dataan application ← from Implement generative AI solutions Implement a prompt flow solutionDesign workflows, tool-augmented flows, and multistep reasoning pipelines ← from Implement generative AI solutionsPrompt flow is no longer named; workflows, tool-augmented flows and multistep reasoning pipelines replace it.- Evaluate models and
flowsapps, including detecting fabrications, relevance, quality, and safety ← from Implement generative AI solutionsEvaluation now names fabrications, relevance, quality and safety. - Integrate generative workflows into applications by using Foundry SDKs and connectors ← one of 2 skills that replace 4 skills of AI-102:
Install and utilize the appropriate SDKs and APIs(from Plan and manage an Azure AI solution)Determine a default endpoint for a service(from Plan and manage an Azure AI solution)Integrate your project into an application with Microsoft Foundry SDK(from Implement generative AI solutions)Integrate Azure OpenAI into your own application(from Implement generative AI solutions)
- Configure an application to connect to a Foundry project ← part of the same change as the skill above
Build agents by using Foundry
Understand the role and use cases of anDefine agent roles, goals, conversation-tracking approach, and tool schemas ← from Implement an agentic solutionUnderstanding agent roles and use cases becomes defining an agent's role, goals, conversation tracking and tool schemas.- Build agents that integrate retrieval, function-calling, and conversation memory ← merged from 2 skills of AI-102:
Create an agent with the Microsoft Foundry Agent Service(from Implement an agentic solution)Implement complex agents with Microsoft Agent Framework(from Implement an agentic solution)
- Integrate agent tools, including APIs, knowledge stores, search, content understanding, and custom functions
- Implement orchestrated multi-agent solutions ← one of 2 skills split from 1 skill of AI-102:
Implement complex workflows including orchestration for a multi-agent solution, multiple users, and autonomous capabilities(from Implement an agentic solution)
- Build autonomous or semiautonomous workflows with safeguards and approval flow controls ← part of the same change as the skill above
Test, optimize and deploy anIntegrate monitoring into deployed agents, evaluate agent behavior, and perform error analysis ← from Implement an agentic solutionTesting and optimization become monitoring, evaluation and error analysis; deployment is covered by model and agent deployment configuration.
Optimize and operationalize generative AI systems
- Tune generation behavior, such as prompt engineering and adjusting model parameters ← merged from 3 skills of AI-102:
Utilize prompt templates in your generative AI solution(from Implement generative AI solutions)Configure parameters to control generative behavior(from Implement generative AI solutions)Apply prompt engineering techniques to improve responses(from Implement generative AI solutions)
- Implement model
reflectionreflection, chain-of-thought evaluations, and self-critique loops ← from Implement generative AI solutionsModel reflection remains and adds chain-of-thought evaluations and self-critique loops. - Set up observability by implementing tracing, token analytics, safety signals, and latency breakdowns ← one of 2 skills that replace 3 skills of AI-102:
Monitor an Azure AI resource(from Plan and manage an Azure AI solution)Configure model monitoring and diagnostic settings, including performance and resource consumption(from Implement generative AI solutions)Enable tracing and collect feedback(from Implement generative AI solutions)
Implement orchestration ofOrchestrate multiplegenerative AI modelsmodels, flows, or hybrid LLM and rules engines ← from Implement generative AI solutions
Implement computer vision solutions (10–15%)
Design and implement image- and video-generation solutions
Use the DALL-E model to generateImplement a solution that generates images from text prompts and reference media ← from Implement generative AI solutionsDALL-E is no longer named; image generation now also accepts reference media.- Implement a solution that generates videos from text prompts and reference media
- Configure image-editing workflows, including inpainting, mask-based edits, and prompt-driven modifications
- Implement workflows to edit generated videos
- Select and apply appropriate generation and editing controls provided by the platform
Design and implement multimodal understanding workflows
- Build a solution that analyzes visual context by using multimodal models ← one of 2 skills that replace 4 skills of AI-102:
Deploy the appropriate generative AI model for your use case(from Implement generative AI solutions)Select and deploy an Azure OpenAI model(from Implement generative AI solutions)Submit prompts to generate code and natural language responses(from Implement generative AI solutions)Use large multimodal models in Azure OpenAI(from Implement generative AI solutions)
- Configure apps to produce concise or detailed captions for single or multiple images
- Implement a solution that enables question-answering grounded in visual evidence
- Configure generation of alt-text and extended image descriptions aligned to accessibility guidelines
- Implement visual understanding by configuring Azure Content Understanding in Foundry Tools to extract visual characteristics ← merged from 3 skills of AI-102:
Select visual features to meet image processing requirements(from Analyze images)Include image analysis features in an image processing request(from Analyze images)Interpret image processing responses(from Analyze images)
Use Azure AI Video IndexerImplement video analysis workflows toextract insights from aprocess and interpret videoor live streamsegments ← from Analyze videosVideo Indexer is no longer named; video analysis becomes a workflow for video segments.- Configure single-task and pro-mode Content Understanding pipelines
- Implement solutions that identify objects, components, or regions within images or video ← merged from 2 skills of AI-102:
Detect objects in images and generate image tags(from Analyze images)Use Azure Vision in Foundry Tools Spatial Analysis to detect presence and movement of people in video(from Analyze videos)
Implement responsible AI for multimodal content
- Implement filters to classify unsafe or disallowed visual content
- Detect and mitigate indirect prompt injection by using embedded text in images
- Enforce visual policy rules, such as applying watermarks, flagging prohibited symbols, upholding brand usage requirements, and detecting potentially inappropriate content
Analyze images
Select visual features to meet image processing requirements→ merged with 2 other skills into one skill in Design and implement multimodal understanding workflowsDetect objects in images and generate image tags→ merged with 1 other skill into one skill in Design and implement multimodal understanding workflowsInclude image analysis features in an image processing request→ merged with 2 other skills into one skill in Design and implement multimodal understanding workflowsInterpret image processing responses→ merged with 2 other skills into one skill in Design and implement multimodal understanding workflowsExtract text from images using Azure Vision in Foundry Tools→ regrouped with 4 other skills into 2 skills in Implement information extraction solutionsConvert handwritten text using Azure Vision in Foundry Tools→ regrouped with 4 other skills into 2 skills in Implement information extraction solutions
Implement custom vision models
Choose between image classification and object detection modelsLabel imagesTrain a custom image model, including image classification and object detectionEvaluate custom vision model metricsPublish a custom vision modelConsume a custom vision modelBuild a custom vision model code first
Analyze videos
Use Azure AI Video Indexer to extract insights from a video or live stream→ now in Design and implement multimodal understanding workflowsUse Azure Vision in Foundry Tools Spatial Analysis to detect presence and movement of people in video→ merged with 1 other skill into one skill in Design and implement multimodal understanding workflows
Implement natural language processing text analysis solutions (15–20% 10–15%)
Apply language model text analysis
Extract key phrasesImplement solutions to extract entities, topics, summaries, andentitiesstructured JSON outputs by using generative prompting and Foundry Tools ← from Analyze and translate textEntity extraction remains; key phrases become topics, summaries and structured JSON from generative prompting and Foundry Tools.- Configure detection of sentiment, tone, safety issues, and sensitive content ← merged from 2 skills of AI-102:
Determine sentiment of text(from Analyze and translate text)Detect personally identifiable information (PII) in text(from Analyze and translate text)
TranslateBuild solutions that translate textand documentsby usingtheAzure Translator in Foundry Toolsserviceor LLM-powered translation flows ← from Analyze and translate textTranslator remains, LLM-powered translation is added, and document translation is no longer named.- Customize language model outputs for domain tasks, such as compliance summarization and domain extraction
Implement speech solutions
- Implement
text-to-speechworkflows to convert speech to text andspeech-to-text using Azure Speech in Foundry Toolstext to speech for agentic interactions ← from Process and translate speechSpeech conversion is now framed for agentic interactions. - Integrate speech as an agent modality, including custom speech models ← merged from 2 skills of AI-102:
Integrate generative AI speaking capabilities in an application(from Process and translate speech)Implement custom speech solutions with Azure Speech in Foundry Tools(from Process and translate speech)
- Enable multimodal reasoning from audio inputs
- Translate
speech-to-speech and speech-to-textspeech into other languages by usingthe Azure Speech inlanguage models and Foundry Toolsservice← from Process and translate speechSpeech translation now also names language models.
Analyze and translate text
Extract key phrases and entities→ now in Apply language model text analysisDetermine sentiment of text→ merged with 1 other skill into one skill in Apply language model text analysisDetect the language used in textDetect personally identifiable information (PII) in text→ merged with 1 other skill into one skill in Apply language model text analysisTranslate text and documents by using the Azure Translator in Foundry Tools service→ now in Apply language model text analysis
Process and translate speech
Integrate generative AI speaking capabilities in an application→ merged with 1 other skill into one skill in Implement speech solutionsImplement text-to-speech and speech-to-text using Azure Speech in Foundry Tools→ now in Implement speech solutionsImprove text-to-speech by using Speech Synthesis Markup Language (SSML)Implement custom speech solutions with Azure Speech in Foundry Tools→ merged with 1 other skill into one skill in Implement speech solutionsImplement intent and keyword recognition with Azure Speech in Foundry ToolsTranslate speech-to-speech and speech-to-text by using the Azure Speech in Foundry Tools service→ now in Implement speech solutions
Implement custom language models
Create intents, entities, and add utterancesTrain, evaluate, deploy, and test a language understanding modelOptimize, backup, and recover language understanding modelConsume a language model from a client applicationCreate a custom question answering projectAdd question-and-answer pairs and import sources for question answeringTrain, test, and publish a knowledge baseCreate a multi-turn conversationAdd alternate phrasing and chit-chat to a knowledge baseExport a knowledge baseCreate a multi-language question answering solutionImplement custom translation, including training, improving, and publishing a custom model
Implement knowledge mining and information extraction solutions (15–20% 10–15%)
Build retrieval and grounding pipelines
- Ingest and index content, such as documents, images, audio, and video ← one of 2 skills that replace 4 skills of AI-102:
Provision an Azure AI Search resource, create an index, and define a skillset(from Implement an Azure AI Search solution)Create data sources and indexers(from Implement an Azure AI Search solution)Implement custom skills and include them in a skillset(from Implement an Azure AI Search solution)Create and run an indexer(from Implement an Azure AI Search solution)
- Configure semantic search, hybrid search, and vector search for grounding ← merged from 2 skills of AI-102:
Query an index, including syntax, sorting, filtering, and wildcards(from Implement an Azure AI Search solution)Implement semantic and vector store solutions(from Implement an Azure AI Search solution)
- Implement enrichment by using custom or built-in skills for text, images, and layout ← one of 2 skills that replace 4 skills of AI-102:
Provision an Azure AI Search resource, create an index, and define a skillset(from Implement an Azure AI Search solution)Create data sources and indexers(from Implement an Azure AI Search solution)Implement custom skills and include them in a skillset(from Implement an Azure AI Search solution)Create and run an indexer(from Implement an Azure AI Search solution)
- Configure RAG ingestion flow, including documents and using optical character recognition (OCR) ← one of 2 skills that replace 5 skills of AI-102:
Extract text from images using Azure Vision in Foundry Tools(from Implement computer vision solutions)Convert handwritten text using Azure Vision in Foundry Tools(from Implement computer vision solutions)Use prebuilt models to extract data from documents(from Implement an Azure Document Intelligence in Foundry Tools solution)Create an OCR pipeline to extract text from images and documents(from Extract information with Azure Content Understanding in Foundry Tools)Extract entities, tables, and images from documents(from Extract information with Azure Content Understanding in Foundry Tools)
- Connect retrieval pipelines directly to workflows and agent tools
Extract content from documents
- Extract information by using multimodal pipelines that combine OCR, layout analysis, and field extraction ← one of 2 skills that replace 5 skills of AI-102:
Extract text from images using Azure Vision in Foundry Tools(from Implement computer vision solutions)Convert handwritten text using Azure Vision in Foundry Tools(from Implement computer vision solutions)Use prebuilt models to extract data from documents(from Implement an Azure Document Intelligence in Foundry Tools solution)Create an OCR pipeline to extract text from images and documents(from Extract information with Azure Content Understanding in Foundry Tools)Extract entities, tables, and images from documents(from Extract information with Azure Content Understanding in Foundry Tools)
Process and ingest documents, images, videos, and audioProduce clean, grounded representations to use withAzureagents and RAG by using Content Understandingin Foundry Tools← from Extract information with Azure Content Understanding in Foundry ToolsContent Understanding processing of documents, images, video and audio now produces grounded representations for agents and RAG.- Implement analyzers for generating structured or markdown outputs for downstream reasoning by using Content Understanding ← merged from 2 skills of AI-102:
Implement a custom document intelligence model(from Implement an Azure Document Intelligence in Foundry Tools solution)Summarize, classify, and detect attributes of documents(from Extract information with Azure Content Understanding in Foundry Tools)
Implement an Azure AI Search solution
Provision an Azure AI Search resource, create an index, and define a skillset→ regrouped with 3 other skills into 2 skills in Build retrieval and grounding pipelinesCreate data sources and indexers→ regrouped with 3 other skills into 2 skills in Build retrieval and grounding pipelinesImplement custom skills and include them in a skillset→ regrouped with 3 other skills into 2 skills in Build retrieval and grounding pipelinesCreate and run an indexer→ regrouped with 3 other skills into 2 skills in Build retrieval and grounding pipelinesQuery an index, including syntax, sorting, filtering, and wildcards→ merged with 1 other skill into one skill in Build retrieval and grounding pipelinesManage Knowledge Store projections, including file, object, and table projectionsImplement semantic and vector store solutions→ merged with 1 other skill into one skill in Build retrieval and grounding pipelines
Implement an Azure Document Intelligence in Foundry Tools solution
Provision a Document Intelligence resourceUse prebuilt models to extract data from documents→ regrouped with 4 other skills into 2 skills in Build retrieval and grounding pipelines and Extract content from documentsImplement a custom document intelligence model→ merged with 1 other skill into one skill in Extract content from documentsTrain, test, and publish a custom document intelligence modelCreate a composed document intelligence model
Extract information with Azure Content Understanding in Foundry Tools
Create an OCR pipeline to extract text from images and documents→ regrouped with 4 other skills into 2 skills in Build retrieval and grounding pipelines and Extract content from documentsSummarize, classify, and detect attributes of documents→ merged with 1 other skill into one skill in Extract content from documentsExtract entities, tables, and images from documents→ regrouped with 4 other skills into 2 skills in Build retrieval and grounding pipelines and Extract content from documentsProcess and ingest documents, images, videos, and audio with Azure Content Understanding in Foundry Tools→ now in Extract content from documents
Implement generative AI solutions (15–20%)
Build generative AI solutions with Microsoft Foundry
Plan and prepare for a generative AI solution→ merged with 4 other skills into one skill in Plan and manage an Azure AI solutionDeploy a hub, project, and necessary resources with Microsoft Foundry→ merged with 4 other skills into one skill in Plan and manage an Azure AI solutionDeploy the appropriate generative AI model for your use case→ regrouped with 3 other skills into 2 skills in Implement generative AI and agentic solutions and Implement computer vision solutionsImplement a prompt flow solution→ now in Implement generative AI and agentic solutionsImplement a RAG pattern by grounding a model in your data→ now in Implement generative AI and agentic solutionsEvaluate models and flows→ now in Implement generative AI and agentic solutionsIntegrate your project into an application with Microsoft Foundry SDK→ regrouped with 3 other skills into 2 skills in Implement generative AI and agentic solutionsUtilize prompt templates in your generative AI solution→ merged with 2 other skills into one skill in Implement generative AI and agentic solutions
Use Azure OpenAI in Foundry Models to generate content
Provision an Azure OpenAI in Foundry Models resource→ merged with 4 other skills into one skill in Plan and manage an Azure AI solutionSelect and deploy an Azure OpenAI model→ regrouped with 3 other skills into 2 skills in Implement generative AI and agentic solutions and Implement computer vision solutionsSubmit prompts to generate code and natural language responses→ regrouped with 3 other skills into 2 skills in Implement generative AI and agentic solutions and Implement computer vision solutionsUse the DALL-E model to generate images→ now in Implement computer vision solutionsIntegrate Azure OpenAI into your own application→ regrouped with 3 other skills into 2 skills in Implement generative AI and agentic solutionsUse large multimodal models in Azure OpenAI→ regrouped with 3 other skills into 2 skills in Implement generative AI and agentic solutions and Implement computer vision solutions
Optimize and operationalize a generative AI solution
Configure parameters to control generative behavior→ merged with 2 other skills into one skill in Implement generative AI and agentic solutionsConfigure model monitoring and diagnostic settings, including performance and resource consumption→ regrouped with 2 other skills into 2 skills in Plan and manage an Azure AI solution and Implement generative AI and agentic solutionsOptimize and manage resources for deployment, including scalability and foundational model updates→ merged with 1 other skill into one skill in Plan and manage an Azure AI solutionEnable tracing and collect feedback→ regrouped with 2 other skills into 2 skills in Plan and manage an Azure AI solution and Implement generative AI and agentic solutionsImplement model reflection→ now in Implement generative AI and agentic solutionsDeploy containers for use on local and edge devicesImplement orchestration of multiple generative AI models→ now in Implement generative AI and agentic solutionsApply prompt engineering techniques to improve responses→ merged with 2 other skills into one skill in Implement generative AI and agentic solutionsFine-tune a generative model
Implement an agentic solution (5–10%)
Create custom agents
Understand the role and use cases of an agent→ now in Implement generative AI and agentic solutionsConfigure the necessary resources to build an agent→ merged with 4 other skills into one skill in Plan and manage an Azure AI solutionCreate an agent with the Microsoft Foundry Agent Service→ merged with 1 other skill into one skill in Implement generative AI and agentic solutionsImplement complex agents with Microsoft Agent Framework→ merged with 1 other skill into one skill in Implement generative AI and agentic solutionsImplement complex workflows including orchestration for a multi-agent solution, multiple users, and autonomous capabilities→ split into 2 skills in Implement generative AI and agentic solutionsTest, optimize and deploy an agent→ now in Implement generative AI and agentic solutions
Current
Microsoft's page labels this version "as of April 16, 2026".
Skills measured as of April 16, 2026
Audience profile
As a candidate for this Microsoft Certification, you’re an Azure AI engineer who builds, manages, and deploys agents and AI solutions that take advantage of Microsoft Foundry.
For this exam, you should have experience developing apps by using Python, and you need to be familiar with the capabilities of general AI, generative AI, and Azure services.
Your responsibilities include:
- Planning and managing Azure AI solutions.
- Implementing generative AI and agentic solutions.
- Implementing computer vision solutions.
- Implementing text analysis solutions.
- Implementing information extraction solutions.
In this role, you collaborate with business stakeholders, solution architects, data scientists, DevOps engineers, and cloud security engineers to design, implement, and maintain AI solutions.
Skills at a glance
- Plan and manage an Azure AI solution (25–30%)
- Implement generative AI and agentic solutions (30–35%)
- Implement computer vision solutions (10–15%)
- Implement text analysis solutions (10–15%)
- Implement information extraction solutions (10–15%)
Plan and manage an Azure AI solution (25–30%)
Choose the appropriate Foundry services for generative AI and agents
- Choose an appropriate model for each task, including large language models (LLMs), small language models, multimodal models, and Foundry Tools
- Choose the appropriate Foundry services for generative tasks, grounding, vector search, agent workflows, or multimodal processing
- Choose an appropriate method for retrieval and indexing
- Choose appropriate memory, tool, and knowledge integration services for agent solutions
Set up AI solutions in Foundry
- Design Azure infrastructure for AI apps and agent-based solutions
- Choose appropriate deployment options
- Configure model and agent deployments
- Integrate Foundry projects with continuous integration and continuous deployment (CI/CD) pipelines
Manage, monitor, and secure AI systems
- Manage quotas, scaling, rate limits, and cost footprints for model and agent workloads
- Monitor model performance, drift, safety events, and grounding quality
- Monitor data ingestion quality, search index health, and relevance performance
- Configure security, including managed identity, private networking, keyless credentials, and role policies
Implement responsible AI across generative AI and agentic systems
- Configure safety filters, guardrails, risk detection, and content moderation
- Apply responsible AI instrumentation, including evaluators, safety evaluations, and explanation tooling
- Implement auditing through trace logging, provenance metadata, and approval workflows
- Govern agent behavior with oversight modes, constraints, and tool-access controls
Implement generative AI and agentic solutions (30–35%)
Build generative applications by using Foundry
- Deploy and consume LLMs, small models, code models, and multimodal models
- Implement retrieval-augmented generation (RAG) in an application
- Design workflows, tool-augmented flows, and multistep reasoning pipelines
- Evaluate models and apps, including detecting fabrications, relevance, quality, and safety
- Integrate generative workflows into applications by using Foundry SDKs and connectors
- Configure an application to connect to a Foundry project
Build agents by using Foundry
- Define agent roles, goals, conversation-tracking approach, and tool schemas
- Build agents that integrate retrieval, function-calling, and conversation memory
- Integrate agent tools, including APIs, knowledge stores, search, content understanding, and custom functions
- Implement orchestrated multi-agent solutions
- Build autonomous or semiautonomous workflows with safeguards and approval flow controls
- Integrate monitoring into deployed agents, evaluate agent behavior, and perform error analysis
Optimize and operationalize generative AI systems
- Tune generation behavior, such as prompt engineering and adjusting model parameters
- Implement model reflection, chain-of-thought evaluations, and self-critique loops
- Set up observability by implementing tracing, token analytics, safety signals, and latency breakdowns
- Orchestrate multiple models, flows, or hybrid LLM and rules engines
Implement computer vision solutions (10–15%)
Design and implement image- and video-generation solutions
- Implement a solution that generates images from text prompts and reference media
- Implement a solution that generates videos from text prompts and reference media
- Configure image-editing workflows, including inpainting, mask-based edits, and prompt-driven modifications
- Implement workflows to edit generated videos
- Select and apply appropriate generation and editing controls provided by the platform
Design and implement multimodal understanding workflows
- Build a solution that analyzes visual context by using multimodal models
- Configure apps to produce concise or detailed captions for single or multiple images
- Implement a solution that enables question-answering grounded in visual evidence
- Configure generation of alt-text and extended image descriptions aligned to accessibility guidelines
- Implement visual understanding by configuring Azure Content Understanding in Foundry Tools to extract visual characteristics
- Implement video analysis workflows to process and interpret video segments
- Configure single-task and pro-mode Content Understanding pipelines
- Implement solutions that identify objects, components, or regions within images or video
Implement responsible AI for multimodal content
- Implement filters to classify unsafe or disallowed visual content
- Detect and mitigate indirect prompt injection by using embedded text in images
- Enforce visual policy rules, such as applying watermarks, flagging prohibited symbols, upholding brand usage requirements, and detecting potentially inappropriate content
Implement text analysis solutions (10–15%)
Apply language model text analysis
- Implement solutions to extract entities, topics, summaries, and structured JSON outputs by using generative prompting and Foundry Tools
- Configure detection of sentiment, tone, safety issues, and sensitive content
- Build solutions that translate text by using Azure Translator in Foundry Tools or LLM-powered translation flows
- Customize language model outputs for domain tasks, such as compliance summarization and domain extraction
Implement speech solutions
- Implement workflows to convert speech to text and text to speech for agentic interactions
- Integrate speech as an agent modality, including custom speech models
- Enable multimodal reasoning from audio inputs
- Translate speech into other languages by using language models and Foundry Tools
Implement information extraction solutions (10–15%)
Build retrieval and grounding pipelines
- Ingest and index content, such as documents, images, audio, and video
- Configure semantic search, hybrid search, and vector search for grounding
- Implement enrichment by using custom or built-in skills for text, images, and layout
- Configure RAG ingestion flow, including documents and using optical character recognition (OCR)
- Connect retrieval pipelines directly to workflows and agent tools
Extract content from documents
- Extract information by using multimodal pipelines that combine OCR, layout analysis, and field extraction
- Produce clean, grounded representations to use with agents and RAG by using Content Understanding
- Implement analyzers for generating structured or markdown outputs for downstream reasoning by using Content Understanding
History
Every recorded change to this study guide, newest first. We started recording on October 15, 2026; earlier history is not available yet.
No changes recorded yet.
October 15, 2026: first version recorded (skills measured as of April 16, 2026).
5 domains, 14 skill groups, 64 skills. See the Current tab for the full outline.