Foundry Tools
Unit 4 of 8 · Module 1: Plan and prepare to develop AI solutions on Azure
Read this lesson
Off-the-shelf AI
Generative models and agents get most of the attention these days, and they're great at open-ended work. But plenty of jobs are routine, like spotting names in a story or reading a receipt, and for those it's smarter to grab something off the shelf. That's what Foundry Tools are: prebuilt AI services and models inside Microsoft Foundry that your apps can call right away. For routine jobs like these, they usually cost less, and give more predictable results, than asking a generative agent every time.
Azure Language
First, Azure Language, for anything written. Hand it an article and it picks out the people, places, and organizations, reads the sentiment, and takes care of summarizing, so a long piece shrinks to a few lines. It also powers conversational language models that work out what a caller wants, and question answering over your own help pages.
Azure Speech
Next, Azure Speech, for anything spoken. It can turn written text into spoken audio, and turn speech back into text, like a court reporter typing every word. It even handles real-time, live conversations, so an agent can talk with a caller while they're still speaking.
Azure Translator
Azure Translator uses modern language models to carry text across a long list of languages. Picture one breaking story reaching readers in Tokyo, Madrid, and Cairo at the same moment.
Azure Document Intelligence
Azure Document Intelligence reads paperwork for you: it pulls the fields out of invoices, receipts, and forms, like the total, the date, and who sent it. Use a prebuilt model for common documents, or train a custom model on a layout only your business uses.
Azure Content Understanding
Azure Content Understanding takes this further, across every kind of media. Give it forms and documents, images, video, or audio, and it turns them into structured information your app can use, like a table of who spoke when in a recorded meeting.
Connecting to the tools
So how does an app actually use one of these tools? You write a client application that sends its requests to that tool's own endpoint, inside your Microsoft Foundry resource. Each request proves who's asking, either with a key, or with token-based authentication through Microsoft Entra ID. Then your code calls the tool through its APIs, or the SDK for your language. And some tools have a page in the Foundry portal, where you can try them out before writing any code.
A note on names
A quick word on names, because they've changed a lot. Before Foundry Tools, the same offerings were called Azure AI Services, and before that, Azure Cognitive Services. Those older names still pop up in some APIs and SDKs, and some tools can still be created as separate Azure resources. For anything new, though, use the tools that come with your Microsoft Foundry resource.
Facts from Microsoft's documentation: What are Foundry Tools?; What is Microsoft Foundry? (CC BY 4.0), rewritten in Loop's own words and examples. Narration, animation, and quiz by Loop. The coding lab is based on code samples in Microsoft's Azure AI documentation (MicrosoftDocs/azure-ai-docs, MIT License; notice). Microsoft, Azure, and Microsoft Foundry are trademarks of Microsoft. Loop is independent and not affiliated with or endorsed by Microsoft.