How to use Hey Copilot: Voice, Vision, Files, and Meetings

Talking to your computer is no longer science fiction: with the "Hey Copilot" command you can summon Microsoft's assistant and ask it things with your voice. Windows 11, since opening apps from summarizing a document to explaining how to connect headphones, all without touching the keyboard, making the interaction faster and more natural than writing.

In addition to the wake word, the experience is enhanced by Copilot Vision, which "sees" what's on your screen when you allow it, and practical features like file search, screenshots, and device context; all with clear privacy controls and options for managing history and voice.

What is "Hey Copilot" and what do you need to use it?

"Hey Copilot" (in some menus it appears as "Hola Copilot") is the hands-free activation word which wakes up Copilot in Windows 11 and opens a listening mode to have a voice conversation with the assistant.

To make it work you need three things: Windows 11 with the Copilot app installed, to sign in with your Microsoft account to unlock history and extended features, and a working microphone (built-in or external), because without audio input no voice recognition possible.

You don't need a PC labeled Copilot+ PC to use voice commands; those devices offer a better experience thanks to hardware with NPU, but any computer with Windows 11 can use Copilot Voice, while the Recall function is exclusive to Copilot+ PC and It works locally and is encrypted..

There's another useful nuance: the use of Copilot Voice is not limited by the number of queries, although subscribers of Microsoft 365 (Personal, Family, or Premium) receive priority access during peak demand periods, and the system supports a wide variety of languages.

Activate voice command step by step

By default, "Hey Copilot" detection is disabled, so you have to enable it manually within the app itself. Open Copilot from the taskbar or Start menu, maximize its window to see all the menus, and go to Account and Settings, where you'll find the Voice section with the toggle to activate it. Listen to "Hey Copilot" to start a conversation.

Once inside Voice, you can also choose your preferred voice for the assistant's spoken responses, so that, in addition to understanding you, it responds in the tone that is most comfortable for you, since the audio output is configurable and can be adjusted to your liking.

If you're having trouble finding the menu, open the left sidebar of the Copilot app (icon next to the assistant logo) and tap your profile icon to jump to Account and Settings; after activating listening, return to the main screen with the house icon and try saying "Hey Copilot" to check that the greeting and the listening mode.

A hardware note: on a desktop computer you'll probably need an external microphone (either a desktop microphone or the one built into the headset), while on a laptop the integrated microphone is usually sufficient. Make sure it's not muted by any physical button or shortcut on the computer, as Copilot needs to be able to "hear" you in order to... enter listening mode.

Ways to talk to Copilot: wake up by voice, press and hold, and microphone

In addition to the "Hey Copilot" command, you can launch Copilot Voice with system shortcuts: press and hold the Copilot key for 1–2 seconds or use Windows + C to open the voice panel, which plays a short greeting and lets you talk to the assistant hands-free—perfect if you're going to dictate several requests in a row.

If you prefer to do it with a click, tap the microphone icon in the Copilot app and listening will begin; the behavior is the same as with the wake word, including the floating box that doesn't obstruct the full screen and the controls to mute or end the session, with an interface clean and direct.

To end, you can say "Cancel" or "End conversation," close with the X on the panel, or simply stop talking for a few seconds, as voice sessions can automatically disconnect when there is no activity or the time limit is reached. time limit.

If you like to configure details, go to Settings > Account > Copilot keyboard shortcuts to choose whether the key opens the full app or a more compact view, and within the same panel, enable "Press and hold" to use the mode of Press to speak whenever it's more convenient for you.

What can Copilot do when it's listening to you?

Once listening mode is open, you can request actions and responses in natural language: open applications, create an email draft, explain a system setting, or ask for a summary of the document in front of you—all conversationally and with the advantage of having the hands-free for other tasks.

When you use Vision alongside voice, the possibilities expand: by allowing Copilot to see the active window or the entire desktop, the assistant understands the visual context, you step by step guide and suggests improvements or shortcuts, such as rearranging an Excel sheet or rewriting a paragraph in Word with a style more formal or concise.

You can also combine it with searching and reading local files (including those synced with OneDrive on your device) to find, open, and view content of common types such as .docx, .xlsx, .pptx, .txt, .pdf, and .json, making Copilot a universal search engine within the PC.

A quick example: "Find my meal prep file" will bring up something like mealprep.docx or mealprep_plan.xlsx; "Files from last week" will list those created during that period; and "Files PDF on my device" will filter by type, always respecting the permissions you have set in File access settings.

Copilot Vision and Recall: what they are and how they differ

Copilot Vision only works during the session and on the content you share at that moment (an Edge tab, a Windows app, or even the camera from the mobile app); it doesn't save screenshots or audio after the interaction, beyond the chat transcript that remains in your history and that You can delete it whenever you want..

Memories, on the other hand, is a feature exclusive to Copilot+ PCs: it takes encrypted snapshots of the screen at the device level and saves them locally, where they are also analyzed; its idea is to be a private "photographic memory" that you can later access, with dedicated privacy controls and without uploading those images to the cloud.

The key is simple: Vision assists you with what you see and choose to share in that moment; Memories creates a local record searchable on compatible devices, all under your control; both capabilities combine with voice so you can ask for explanations, summaries, or actions based on what's in your screen or in your history.

If you don't want Vision to see anything, simply don't share your screen or close that session; if you don't want Memories, don't enable it or review its privacy settings in your system, because Microsoft clearly separates both concepts so you can decide. when and how do they intervene.

Find and understand your files on your PC

Copilot in Windows can use File Search and File Read permissions to help you find and open documents, and also answer questions about their contents: the assistant can read a PDF, Word, or TXT file and tell you the essentials, quote sections, or convert it into a concrete to-do list.

Go to Settings > Account > File Search and File Reading to turn these capabilities on or off; your control is granular and you can limit it if you prefer the assistant to only see file names, not their content, reinforcing that you are the one setting the tone. adequate level of access.

When you work with OneDrive synced on your own computer, the experience is seamless: it behaves as if everything were local, so you can find what you need without worrying about paths or folders, as Copilot understands the query and returns results. filtered by relevance.

In combination with Vision, the search becomes even more direct: if you have a report open, you can ask "summarize what was decided in this document" and the assistant will contextualize it based on what it sees on the screen, saving you from scrolling page by page. long documents.

Screenshots and device context

If you need to ask Copilot about something that appears on your screen, use the capture option: in the Copilot compositor, tap + and then "Take a screenshot" to select an area or the entire screen; then formulate your question and send it for the assistant to answer. specific fragment.

The "device context" also matters: you can ask "How do I pair Bluetooth headphones on My PC"?" and Copilot will respond with steps tailored to your version of Windows, avoiding generic instructions that don't fit your system, which is particularly useful when The menus change.

This blend of capture and context resolves doubts on the fly: whether it's interpreting an error in an app, comparing figures you have in view, or requesting that a paragraph you're editing be rewritten, Copilot understands the shared space and gives you step-by-step assistance. without deviating from what you're doing.

Remember that, by design, Vision does not store these captures once the session ends; all that remains is the transcript in your history, which you can delete at any time, with the peace of mind that the shared image or audio will not be lost. They keep in Copilot.

Windows shortcuts and "push to speak"

The Copilot key and Windows + C are the fastest routes to open the app or the compact view; in Settings > Account > Copilot keyboard shortcuts you decide whether the shortcut launches the entire app or a lightweight window, ideal for quick queries and keeping Copilot handy without invading your desktop.

The "Push to Talk" mode (press and hold) is activated in that same section and starts Copilot Voice when you hold the corresponding key for 1 to 2 seconds; releasing it ends the listening session, which is very practical if you want granular control of the microphone and to prevent the app from interfering. be attentive to everything There.

If you prefer not to use shortcuts, you can always use the microphone icon within the app; you'll see a sound greeting and a listening indicator, as well as the option to mute and unmute with another click, which is the most direct way to switch between voice and audio. keyboard input.

As an added bonus, the Copilot app for Windows includes features not found elsewhere: in addition to the wake word and Vision, you get file search, screenshots, and device context; at the same time, certain experiences like Copilot Shopping or the Copilot Podcast are not available in the Windows app and are maintained in other apps. Copilot surfaces.

What if you can't speak? Vision also with text

There are situations where speaking is impractical (a silent meeting or a noisy café), so Microsoft is bringing the Vision approach to text interaction as well: you'll be able to type your query and benefit from the analysis of what's on the screen, an improvement that first arrives in the Windows Insider program.

Meanwhile, remember that you can seamlessly switch between them: typing when speaking isn't convenient, and using "Hey Copilot" or the microphone when it's more productive, because the assistant understands both modes and keeps the history unified so you can easily find what you're looking for later. that you have already consulted.

This multimodal approach is what Microsoft is pushing: voice, text, and vision when you grant permission, so that Copilot is useful in any context and doesn't get stuck on a single form of interaction, which is key to gaining traction in the Windows desktop.

As an interesting fact shared by Microsoft itself, those who use voice tend to interact twice as much as those who type, which makes sense because speaking is usually faster than typing and because the assistant, by listening to you, can chain together several actions with less friction.

Copilot in Microsoft Teams meetings

At work, Copilot assists within Teams meetings: it summarizes ideas, notes who said what, suggests actions, and answers questions in real time, during and after the meeting, always subject to the transcription policy and whatever the organizer has set. decided to allow.

If you host the meeting, you can choose: "During and after the meeting" (requires transcription; if you stop transcription, Copilot stops), "Only during the meeting" (uses temporary speech-to-text unless someone activates transcription), or "Off" (no Copilot, no recording, and no transcription), all in line with the confidentiality labels.

Once inside, launch Copilot from the meeting controls and ask it questions: what questions should be asked to move forward, where there are gaps in someone's argument, or to create a table with pros and cons; you can also ask it to list ideas by quoting the transcript and chat, with a button of more suggestions.

If you log in late (for example, 5 minutes late), you'll see a notification to catch up with a summary in the sidebar; and if you're on the desktop app, you can open Copilot in a new window to work in parallel without losing sight of the conversation—very practical for real multitasking.

Long or table-based responses can be exported to Word or Excel, unless the meeting has a confidentiality label prohibiting it or the organizer has disabled copying and forwarding the chat. subtitles Live streams or transcripts; after the meeting, access Copilot from the meeting chat or the Summary tab, provided there is transcript available.

Change the voice you hear

The first time you install Copilot, it asks you to choose a voice; afterwards, you can change it whenever you want: on the web (copilot.com) from the profile icon and the Voice section, on Windows or macOS from Settings > Voice, and on mobile within the app menu, by tapping Voice settings to also adjust the playback speed.

This is useful if you spend many hours with the assistant: finding a voice that is clear and pleasant reduces fatigue and makes dictation sessions smoother, especially if you're multitasking or using Copilot as support while performing other tasks. actions on screen.

Privacy and use of your voice data

Voice data is used to provide the Copilot Voice service, and you can control whether conversations are used for model training; the generated transcripts are managed like the rest of your Copilot chat history, and you have the option to clear that history whenever you want.

With Vision, the assistant only accesses the content you share in that session and doesn't save screenshots or audio; with Memories, if you choose to enable it on a Copilot+ PC, snapshots are saved encrypted and analyzed locally, reinforcing that the "memory" it offers is private, controlled by you, and limited to the device.

If you prefer not to use voice in certain environments, temporarily disable "Hey Copilot" listening and use click-to-talk or text input; the goal is to give you control over when and how the system listens to you, with quick options to stop, mute, or end a conversation. complete and instant.

At the end of each voice session, Copilot can request feedback; taking a few seconds to send it helps improve recognition accuracy and response quality, something that is especially noticeable in accents and noisy environments.

New features on the horizon: Text-based vision, Copilot Actions, and more

Microsoft is adding capabilities in waves through the Microsoft Store in Windows 11: in addition to the arrival of Text Vision, it's preparing Copilot Actions to act directly on your files (organizing photos, extracting text from PDFs, or scheduling a meeting by hovering over an email), with the caveat that in complex interfaces can be wrong.

Another new feature mentioned is Manus, which converts documents into websites from Explorer with a right-click, and the ability to edit video with Filmora without leaving the workflow; these are examples of how Copilot wants to integrate with more everyday tasks and encourage Windows 10 users to make the switch, just as that system loses official support.

With Vision, the proposition is clear: the assistant "sees" what you're looking at to help you with context; you choose what to share, and if you don't want to speak, you can also type, letting Copilot be your copilot on the desktop with both voice and text. text and view.

However, Microsoft insists that everything works with explicit authorization and that you can turn it off whenever you want; it's normal to have doubts, but the activation options, local Memories management, and history control provide... tranquility and transparency.

Quick tips and solutions to common problems

If "Hey Copilot" isn't responding, check that the app is up to date, that the window was maximized when configuring voice, and that the microphone isn't muted by hardware. On a desktop computer, confirm that you've selected the correct input device in both Windows and Copilot, because sometimes the system selects a different one. microphone you don't use.

If there's a lot of background noise, try moving closer to the microphone or using a headset; voice recognition is robust, but like any dictation system, it improves when it receives clean audio, resulting in fewer repetitions and more direct responses. accurate on the first try.

If the voice option doesn't appear, check your regional and language settings, and make sure you're signed in to Copilot; signing in unlocks features like history, image generation, and longer conversations, including the voice interaction.

And if the keyboard shortcut doesn't open what you expect, go to Settings > Account > Copilot keyboard shortcuts to choose whether you want the full app or the smaller "quick view," which is ideal for checking without taking up the entire screen. work screen.

Finally, if you want the complete experience, combine voice, Vision, file search, and device context: with "Hey Copilot" you can activate the assistant at any time, ask it to see what's on your screen, find a document, and close it with "Cancel" without touching anything—a convenient way to save time and reduce unnecessary clicks.

With "Hey Copilot" enabled, Copilot Vision when you allow it, access to your files under specific permissions, and integration into Teams meetings, Microsoft's assistant becomes a versatile companion for everyday use in Windows: fast when you speak, helpful when you type, respectful of your privacy settings, and always at hand with shortcuts or a simple voice command that, frankly, makes your PC feel like a slightly smarter.