1. Which service should I use now?
2. What happened to the older tools?

1. What each of the new services does
2. Which ones are still active, renamed, or retired

Azure AI Services – Previously Cognitive Services

Formerly known as Cognitive Services, this was Microsoft original collection of pre-trained, ready-to-use AI APIs. In 2023, it was renamed to Azure AI Services to align with Microsoft rebranding strategy and to unify all prebuilt AI capabilities under a single banner.

These services are great for quickly adding AI features for Vision, Speech, Language and Decision without training a model yourself. Instead of writing complex ML code, developers can just call an API.

Azure OpenAI Service

Introduced in 2021 as part of the partnership between Microsoft and OpenAI, Azure OpenAI Service gives you access to powerful language models like ChatGPT (GPT-4, GPT-4o) with enterprise level security. It extends foundational AI capabilities originally offered through OpenAI APIs and brings them into the enterprise Azure ecosystem.

Azure OpenAI is integrated with other Microsoft services like Azure AI Studio (for workflow orchestration)

Microsoft Copilot (for embedding in Office tools), making it a central part of Microsoft generative AI platform. Its use cases span across business functions like automating customer support, content generation, code assistance, and enterprise search.

Business Use case

You want to build a chatbot, summarizer, or custom Copilot using GPT. Common business use cases include customer service chatbots, automated support agents, sales assistance tools, HR knowledge bots, marketing content generators, and internal productivity copilots that summarize documents or generate emails. These solutions help reduce human workload, speed up communication, and increase customer satisfaction across industries like retail, banking, healthcare, and logistics.

Azure AI Studio – Previously Azure OpenAI Studio

Introduced in 2024 as the evolution of Azure OpenAI Studio, Azure AI Studio is a no-code/low-code interface designed to simplify building AI-powered solutions. The previous name, Azure OpenAI Studio, is now retired and replaced by Azure AI Studio to reflect a broader vision beyond OpenAI models. It still retains OpenAI integration but now also supports orchestration across Azure AI Services and external enterprise connectors.

Azure AI Studio does not directly integrate with Azure AI Foundry but complements it. While Foundry handles model lifecycle, governance, and large-scale deployments (especially for multi-model environments like Meta, Mistral, or Gemini), AI Studio is focused on app building and orchestration. Both services often work together: models deployed and managed via Foundry can be used in workflows created in AI Studio.

This separation of responsibilities allows businesses to prototype in AI Studio and later scale and govern their models using Azure AI Foundry.

Azure AI Studio extends and integrates features from Azure OpenAI, Azure AI Services, and enterprise connectors. It allows users to build AI workflows visually, combine OpenAI models with enterprise data, and create custom copilots or agents that respond to real-time inputs.

It supports prompt engineering, orchestration logic, evaluation, and deployment all in one place. Over time, Azure AI Studio has absorbed earlier features from tools like QnA Maker (via Azure AI Language) and Azure Cognitive Search integration.

Business Use case

•  HR team can build a bot that answers employee policy questions, or a finance team can automate invoice summaries. Legal departments use it to scan and extract terms from long contracts. Marketing teams can use it to create content drafts or analyze customer feedback. Sales teams benefit by generating lead summaries or drafting personalized emails.
• In healthcare, AI Studio is used for symptom checkers or analyzing patient reports. These tools integrate easily with Microsoft Teams, Power Automate, and enterprise data sources to create smart, context-aware workflows.

Azure Machine Learning (Azure ML)

Azure Machine Learning, also known as Azure ML, has been around since 2018 and remains a core component of Microsoft's AI platform. Unlike other Azure AI services focused on prebuilt or foundation models, Azure ML is designed for building, training, and deploying custom machine learning models, especially using structured/tabular data.

Azure ML supports AutoML (automated model building), responsible AI (fairness, transparency), model explainability, MLOps (CI/CD for ML), and deployment to edge devices.

Business use case

Many businesses across industries are using Azure AI to solve real-world challenges and improve productivity. In retail, companies use AI to personalize shopping experiences and automate customer service.

• Healthcare providers are using Azure AI to summarize patient records, generate discharge notes, and assist in diagnosis. Financial institutions benefit from fraud detection, customer support automation, and investment analysis.
• Manufacturing firms apply AI for predictive maintenance, quality checks, and supply chain optimization. Legal teams use AI to summarize contracts, identify risks, and speed up document review.

Azure AI Foundry

It does not replace any previous single tool but instead extends the AI ecosystem by enabling businesses to manage multiple foundational models not just OpenAI, but also Meta LLaMA, Google Gemma, Mistral, Hugging Face, and custom models from a unified control plane.

Foundry is not a renaming of an old tool—it’s a new platform aimed at model lifecycle management and cross-model orchestration.

Conclusion 

• Use Azure ML to train a model
• Host and manage it via AI Foundry
• Expose it via Azure AI Studio
• Integrate it into apps with Azure OpenAI