Copilot vs Copilot Studio vs Foundry vs GitHub
Four Microsoft products share the word Copilot or sit next to it, and exams test whether you can tell them apart. Microsoft 365 Copilot is the assistant you use inside Word, Excel, Outlook and Teams. Copilot Studio is the low-code tool for building agents. Microsoft Foundry is the developer platform for building AI apps and agents with code and choosing models. GitHub Copilot is the AI coding assistant for software developers. The decision rule: ask who the user is and whether they are consuming, configuring or coding.
The what-is-what table
| Product | What it is | Who works with it | Typical exam phrase |
|---|---|---|---|
| Microsoft 365 Copilot | AI assistant in Microsoft 365 apps and chat, grounded in web data and, with the full licence, in your work data | Any employee | “summarise my meetings”, “draft from my files” |
| Copilot Studio | Graphical, low-code studio for building and managing agents and workflows, published to channels such as Teams and websites | Makers, business analysts, citizen developers | “low-code agent”, “topics”, “connectors”, “publish to Teams” |
| Microsoft Foundry | Platform to build agents, models and apps, with a model catalog, evaluation, tracing and governance | Developers and AI engineers | “choose and deploy a model”, “hosted agent”, “evaluate”, “content filter” |
| GitHub Copilot | AI assistant for writing, understanding and shipping code, with completion, chat, agent features and code review | Software developers | “suggest code in the IDE”, “review a pull request” |
Microsoft Learn describes Copilot Studio as a graphical, low-code studio for building and managing AI-powered agents and workflows, and Foundry as a platform that unifies agents, models and tools under a single management grouping with tracing, monitoring and evaluations.
Names to treat with care
Microsoft is renaming and repackaging in this area, so exam wording and documentation can lag each other.
- Foundry. The platform was previously called Azure AI Studio and then Azure AI Foundry; Microsoft Learn now says Microsoft Foundry and lists Azure AI Services as now Foundry Tools. Old material may use either name.
- Copilot naming. Microsoft’s overview now distinguishes three licence experiences: Copilot Chat (Basic), Microsoft 365 Copilot (Basic) and Microsoft 365 Copilot (Premium). It says the Microsoft 365 Copilot app is now called Microsoft Copilot and the Microsoft 365 Copilot Chat app is now called Microsoft Copilot Chat. The Premium experience grounds answers in work data through Microsoft Graph and Work IQ (Work IQ can be turned off); Copilot Chat works with web data and only uses work content you provide with the prompt, unless a pay-as-you-go agent has access. Check the current licensing page before a Copilot exam.
- Copilot Studio harnesses. The docs describe a choice of engine, a GitHub Copilot harness, a standard harness and a Copilot chat harness, which affects capability and billing. Parts of the experience, such as app creation and building from natural language, are labelled preview.
- Microsoft’s overview lists Security Copilot, GitHub Copilot and Copilot Studio as independently licensed products that do not need a Microsoft 365 Copilot (Premium) licence.
A worked example: one request, three answers
A company asks for an assistant that answers HR questions.
- Employees only need help with their own work (summarise a policy PDF they open, draft a reply): Microsoft 365 Copilot or Copilot Chat. Nothing is built.
- HR wants a custom agent that answers from the HR handbook, calls a leave-balance connector and is published to Teams, built by an analyst without writing code: Copilot Studio. This is the extend step, building an agent that adds knowledge and actions to the Copilot experience.
- The product team wants a customer-facing assistant inside their own application, with a chosen model, custom orchestration, evaluation runs and network controls: Microsoft Foundry, with retrieval from a service such as Azure AI Search (see RAG explained).
- Developers building any of the above use GitHub Copilot in their editor to write the code. It is a tool for the builders, not for the end user.
AB-731 and AB-100 both ask when to build, buy or extend; a sensible default is to use what exists first, extend second and build last. Foundry sits at the build end. Agents as a concept are in AI agents explained.
Which exams test this
| Exam | How it appears | On this site |
|---|---|---|
| AB-731 | Microsoft’s AI apps and services domain (35–40%): differences between Copilot versions, Copilot Studio and Foundry capabilities, mapping use cases to them, and when to build, buy or extend | Microsoft AI apps and services |
| AB-100 | Plan AI-powered business solutions (25–30%): design multi-agent solutions across Microsoft 365 Copilot, Copilot Studio and Foundry, set rules for components built in each, and decide when to build custom agents or extend Microsoft 365 Copilot | Plan AI solutions |
| AB-620 | Integrate and extend agents in Copilot Studio (40–45%): bring a Foundry agent into a Copilot Studio solution and use the Foundry model catalog and Azure AI Search with Foundry from Copilot Studio | Integrate and extend |
| AB-730 | Generative AI fundamentals domain (25–30%): when to use Microsoft 365 Copilot Chat, agents or Copilot Cowork, and how Copilot experiences differ; the Copilot side only, not Copilot Studio or Foundry | Generative AI fundamentals |
| AB-900 | Administrative tasks for Copilot and agents (25–30%): compare built-in Copilot capabilities with agents, Copilot licence models and use cases for custom agents; the Copilot side only | Copilot and agent admin |
| AI-901 | Implement AI solutions by using Microsoft Foundry (55–60%): deploy models and build apps and agents in Foundry; Foundry is tested on its own, not compared with Copilot | Foundry implementation |
| AI-103 | Plan and manage an Azure AI solution (25–30%): choose the right Foundry services and models, then build apps and agents on Foundry; Foundry only, no Copilot comparison | Plan and manage |
| DP-800 | Design and develop database solutions domain: enable GitHub Copilot and Copilot in Fabric, configure model and MCP options and instruction files; GitHub Copilot only, a minor part of this topic | Design and develop |
How exam questions are built around these products
The question gives a requester and a skill level, and each wrong option is a product aimed at someone else.
- “Without writing code”, “business users”, “low-code”: Copilot Studio. Not Foundry.
- “Choose a model”, “deploy and evaluate”, “full control”, “custom orchestration in our own app”: Foundry.
- “Based on emails, meetings and files the user can already access”: Microsoft 365 Copilot with Microsoft Graph, and it respects the user’s permissions, which is why oversharing in SharePoint becomes a Copilot problem.
- “Developers want suggestions while typing code”: GitHub Copilot.
- “Add one data source to Copilot”: extend with an agent or connector, do not build from scratch.
- Security or compliance wording such as sensitivity labels and data loss prevention applies to Copilot’s data access; see sensitivity labels explained and DLP explained.
- Product-name distractors. If an option names a retired product name, the underlying capability is usually still the point; read for the role, not the label.
Two practice questions
Question 1. A procurement team wants an agent that answers supplier questions from a shared policy library and creates a purchase request through an existing connector. The team has no developers and wants to publish the agent in Microsoft Teams. Which product fits best?
- A. Microsoft Foundry with a hosted agent
- B. GitHub Copilot
- C. Microsoft Copilot Studio
- D. Microsoft 365 Copilot used as is
Show answer
Answer: C
Copilot Studio is the low-code studio for building agents with knowledge sources, connectors and tools and publishing them to channels such as Teams, which matches a team without developers. Microsoft Foundry is for developers who need code-level control and model choice. GitHub Copilot assists developers in writing code. Microsoft 365 Copilot is for using an assistant in the apps, not building a custom agent.
Question 2. A product team is building a customer-facing assistant into its own web application. It must compare several models from different providers, run evaluations before release, and apply content filters and role-based access. Which Microsoft product is the best fit for the platform?
- A. Microsoft 365 Copilot
- B. Microsoft Foundry
- C. Microsoft Copilot Studio
- D. GitHub Copilot
Show answer
Answer: B
Microsoft Foundry provides a model catalog, agents, evaluations, tracing and governance controls such as content filters and role-based access for developers building their own AI apps. Copilot Studio is aimed at low-code makers working within the Copilot ecosystem. Microsoft 365 Copilot is an assistant inside Microsoft 365 apps. GitHub Copilot helps write code but is not the platform that hosts and governs the assistant.
Go deeper
- What is Microsoft Copilot?: the experiences, licences and data grounding.
- Microsoft Copilot Studio overview: agents, workflows and harnesses.
- What is Microsoft Foundry?: capabilities and the rename history.
- About GitHub Copilot: features and plans.
- Related: fine-tuning vs RAG vs prompting, responsible AI principles.