Azure AI is a portfolio, not one interchangeable product: Microsoft offers task-specific tools for jobs such as translation and document extraction, model access for generative AI, search for retrieving private content, agent tooling, and Azure Machine Learning for custom models. Start with the output your application needs, then choose the service designed to produce it.
What are Microsoft Azure AI services?
Microsoft groups its prebuilt and customizable AI APIs and models under Foundry Tools. They add focused capabilities—such as language processing, speech, translation, vision, search, and content safety—to applications. Other parts of the Azure AI portfolio provide foundation models, agents, and custom machine learning. Each solves a different part of an application rather than acting as a universal AI service.
As an Amazon Associate I earn from qualifying purchases.
Microsoft’s current overview lists Speech, Translator, Language, Content Understanding, Document Intelligence, Vision, Azure AI Search, Content Safety, Custom Vision, and Immersive Reader as Foundry Tools. Some older Azure AI and Cognitive Services documentation uses different names, so check the current product page for the exact name and placement. Microsoft’s AI services overview describes the tool portfolio.
Which Azure AI service fits the task?
| What you need to do | Good starting point | Why it fits |
|---|---|---|
| Analyze text for sentiment, key phrases, entities, summaries, classification, language, question answering, or conversational intent | Azure Language | It provides targeted natural-language capabilities. For document retrieval, consider Azure AI Search; for translation, use Translator. Microsoft’s Language overview distinguishes these uses. |
| Translate text or documents | Azure Translator | It supports real-time text translation, batch or single-file document translation, and custom translation for specialized terminology. Microsoft’s Translator overview describes the options. |
| Extract fields, tables, or structure from forms and documents | Azure Document Intelligence | It offers prebuilt document models as well as custom model options. Microsoft’s Document Intelligence overview explains the service. |
| Extract schema-defined fields from varied media or documents using natural-language descriptions | Azure Content Understanding | Consider it when a suitable prebuilt Document Intelligence model is unavailable or when a workflow needs confidence scores, grounding, or RAG-ready Markdown. See Microsoft’s Content Understanding overview. |
| Transcribe audio, synthesize speech, translate speech, or build speech interaction | Azure Speech | The service includes speech-to-text, text-to-speech, translation, and speaker recognition. Microsoft’s Speech overview describes its capabilities. |
| Analyze image or video content | Azure Vision; consider Content Understanding for broader media extraction | Microsoft’s image and video processing guidance covers Vision and Content Understanding. Review the AI services overview for current portfolio placement. |
| Search a document collection or retrieve relevant material for a conversational application | Azure AI Search | It indexes and retrieves relevant content. Microsoft’s guidance includes it in retrieval-augmented generation (RAG) workflows; Language guidance points to it for searching documents by chat. See the AI services overview and Language overview. |
| Check user-generated or AI-generated text and images for harmful or unwanted content | Content Safety | It provides content checks. Microsoft describes it within its current AI services portfolio; confirm its current product placement and availability for your deployment. See the AI services overview. |
| Generate, summarize, reason over, or understand content with a foundation model | Azure OpenAI or another suitable Foundry Model | Microsoft provides managed access to models and a broader model catalog. Choose a specific model based on your task and check its current regional availability. See Microsoft Foundry documentation. |
| Build an agent that uses a model plus tools or knowledge | Foundry Agent Service | It hosts agents connected to a model and optionally to custom knowledge stores or APIs. See Microsoft Foundry documentation. |
| Train a bespoke model or customize beyond what a prebuilt tool supports | Azure Machine Learning | It is the custom machine-learning path when prebuilt capabilities do not meet the requirement. It typically calls for more machine-learning expertise than using a prebuilt API. See Microsoft Foundry documentation and the AI services overview. |
How to choose the right starting point
- Define the output. Be specific: extracted document fields, translated text, transcribed audio, image labels, sentiment, grounded answers over private content, or generated content. The input alone—such as “documents” or “audio”—does not determine the service.
- Match the job to a prebuilt tool first. If a Foundry Tool documents the capability you need, start there. Microsoft notes that many projects can use prebuilt models or SaaS capabilities, with customization available for some services. Building and operating a custom model adds work and is not automatically necessary.
- Separate retrieval from generation. Azure AI Search indexes and retrieves relevant material; a language model generates or reasons over content. A grounded-answer application may need both. Evaluate whether retrieval finds the right sources and whether the model uses them appropriately; do not assume a model alone searches your private document collection.
- Choose Azure Machine Learning when the prebuilt route falls short. A bespoke model or training approach can better fit a specialized requirement, but weigh that benefit against the additional data, expertise, operations, and governance it entails.
- Check deployment details before implementation. Confirm that the specific service, feature, and model are available in your intended region, then verify pricing, quota, API version, data handling, security controls, and retirement notices. Portfolio descriptions do not establish those details for every deployment.
How to decide when services overlap
When several options could fit, compare them against the workload rather than choosing by product name alone:
- Input and output: Is the input text, a document, audio, or image/video? Do you need classification, extraction, translation, retrieval, or generation?
- How much customization you need: Does a prebuilt service suffice, can it be customized, or do you need to train a model?
- Retrieval and grounding: Must an answer draw on a private or changing collection of material, and how will you assess whether relevant content is found?
- Content and deployment fit: Check supported languages and file formats, regional availability, and data residency for the service and model you intend to use.
- Operational constraints: Compare expected volume, latency, and cost, along with API lifecycle, identity, network isolation, safety, and monitoring requirements.
The first questions narrow the service family. Pricing, regional availability, quotas, and model availability require current checks for the exact service and deployment; they cannot be inferred from a portfolio overview.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How the main Azure AI building blocks fit together
A single application may use multiple services because the parts have separate jobs. For example, a document-question-answering workflow could use Document Intelligence to extract structure from uploaded files, Azure AI Search to index and retrieve relevant passages, a foundation model to draft a response, and Content Safety to check content. Which pieces are appropriate depends on the workflow; a generative model does not eliminate the need to select retrieval, safety, and evaluation components.
Rank #2
Microsoft Foundry is the current unifying platform terminology in the reviewed documentation for models, agents, and tools. Older Azure AI and Cognitive Services materials may still use earlier names, so confirm the current service page before designing around a feature. Microsoft’s Foundry overview describes the platform.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Quick Recap
Rank #4
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
