Copilot is an AI assistant. The name covers two quite different products: GitHub Copilot, which suggests code as you type, and Microsoft Copilot, which works across Windows, Word, Excel, Outlook and Teams. Both use large language models; they do very different jobs.
What is Copilot?
“Copilot” is now a family name rather than one product, which is the first thing to sort out:
| Product | What it is |
|---|---|
| GitHub Copilot | An AI coding assistant that suggests code inside your editor |
| Microsoft Copilot | A general AI assistant in Windows, Edge and on the web |
| Microsoft 365 Copilot | The assistant built into Word, Excel, PowerPoint, Outlook and Teams |
| Copilot in specific apps | Assistants inside Dynamics, Power BI, Security and other Microsoft tools |
They share a brand and an underlying approach — large language models trained on vast amounts of text and code — and beyond that they are separate tools with separate subscriptions.
GitHub Copilot watches what you are writing and suggests the next line, the next block, or a whole function. It works in editors including VS Code and JetBrains, and it can also answer questions in a chat panel about the code in front of you.
Microsoft Copilot answers questions, drafts text, summarises documents and works across the Office applications — writing a first draft in Word, building a formula in Excel, summarising a Teams meeting you missed, or drafting a reply in Outlook.
The pattern in both is the same: it produces a first draft quickly, and you check it. That is where the value is, and also where the risk is.
What is Copilot used for?
GitHub Copilot:
- Autocompleting code, from a line to an entire function.
- Writing boilerplate — the repetitive scaffolding that takes time and no thought.
- Writing tests for existing code.
- Explaining unfamiliar code, which is genuinely useful when inheriting a codebase.
- Translating between languages, converting a script from one language to another.
- Writing documentation and comments.
- Suggesting fixes for errors.
- Working in an unfamiliar language or framework, where it shortens the syntax-lookup loop considerably.
Microsoft Copilot and Microsoft 365 Copilot:
- Drafting documents, emails and reports from a prompt.
- Summarising long documents, email threads and meetings.
- Building Excel formulas, and analysing and charting data.
- Creating presentation drafts from a document.
- Catching up on missed Teams meetings, with a summary and action points.
- Answering questions and searching the web.
- Rewriting text — shortening, changing tone, making it clearer.
- Generating images, in some versions.
What Copilot is not: a replacement for knowing what you are doing. It is confidently wrong often enough that unchecked output is a liability, and that is true of code and prose alike.
Common tasks you can do with Copilot
- Write a function from a comment describing what you want.
- Generate a unit test suite for code you already have.
- Summarise a 40-page document into a page.
- Turn meeting notes into an action list.
- Build a complex Excel formula by describing the result you want.
- Draft a first version of an email, then edit it into your own voice.
- Explain a regular expression or a block of unfamiliar code.
- Convert a Python script into another language.
- Produce a presentation outline from a report.
Who Copilot is useful for
- Developers, particularly for boilerplate, tests and unfamiliar languages.
- Students learning to code, with an important caveat — leaning on it too early can prevent you learning to reason about code yourself.
- Office workers drafting, summarising and reporting.
- Managers catching up on meetings and threads.
- Analysts, for Excel formulas and first-pass data work.
- People who write a lot and want a faster first draft.
- Anyone facing a blank page, which is where it consistently helps most.
Who benefits least: anyone who cannot evaluate the output. If you cannot tell whether the code is right or the summary is accurate, the tool is producing confident work you cannot check — which is worse than no tool.
When to choose Copilot instead of a related tool
| Tool | Best for |
|---|---|
| GitHub Copilot | Code suggestions inside your editor, in context |
| Microsoft 365 Copilot | Drafting and summarising inside Office and Teams |
| ChatGPT, Claude and similar | General-purpose questions, reasoning and writing |
| Traditional autocomplete | Deterministic, predictable completions |
| Search engines | Finding sources you can verify |
Choose Copilot when the work sits inside your editor or your Office documents, where its access to that context is the advantage.
Several things are worth knowing before relying on it:
- Verify everything. It generates plausible output, not verified output — inaccurate code and confidently wrong summaries both happen.
- Understand what it can see. Microsoft 365 Copilot works across your organisation’s documents, which is powerful and means access permissions matter a great deal.
- Check your organisation’s policy before putting confidential or personal information into any AI assistant.
- Review generated code for security, not just correctness. Copilot suggests patterns from its training data, including insecure ones.
- Understand the licensing position on generated code, which remains an evolving area.
- Do not let it substitute for learning, if you are learning.
See the AI and automation hub, including Python and streaming data.
Related software and productivity guides
Frequently asked questions
What is Copilot used for? Depends which one. GitHub Copilot suggests code as you type. Microsoft Copilot drafts, summarises and answers questions across Windows and Office.
Are GitHub Copilot and Microsoft Copilot the same? No. They share a brand and an approach but are separate products with separate subscriptions and different purposes.
Is it free? Both have paid tiers. Microsoft Copilot has a free tier for general use; GitHub Copilot is free for verified students, teachers and some open-source maintainers.
Can I trust the code it writes? Not without checking. It produces plausible code that is sometimes wrong and sometimes insecure. Review it as you would a colleague’s first draft.
Will it replace programmers? It changes what programmers spend time on. Someone still has to decide what to build, judge whether the output is right, and fix it when it is not.
Is my data private? That depends on your plan and your organisation’s configuration. Check your organisation’s policy before entering confidential information.
Does it work in Excel? Yes, in Microsoft 365 Copilot — building formulas, analysing data and creating charts from a description.
Should I use it while learning to code? With care. It is useful for explanations, but leaning on its suggestions too early can stop you learning to reason about code yourself.