Windows 11 updates for Copilot+ PCs equipped with neural processing units (NPUs) introduce contextual on-screen intelligence designed to eliminate repetitive copy-pasting, searching, and photo retouching. One of the flagship local AI capabilities is Click to Do. Operating as an interactive intelligence overlay, Click to Do analyzes whatever is currently displayed on your screen—whether inside a PDF, an image editor, a web browser, or a video call—and surfaces one-click actions tailored specifically to the highlighted content.
Hardware Requirements and System Prerequisites
Because Click to Do relies on local multimodal AI models running in real time across Windows 11 without cloud processing delays, specific hardware conditions must be met:
- Copilot+ PC Architecture: A device powered by an approved processor with an integrated NPU delivering at least 40 TOPS (Tera Operations Per Second), such as Snapdragon X Elite, Snapdragon X Plus, Intel Core Ultra 200V series, or AMD Ryzen AI 300 series.
- Operating System: Windows 11 version 24H2 or newer with the latest feature experience packs installed.
- Windows Recall Architecture: Click to Do leverages the underlying secure snapshot analysis pipeline while maintaining strict on-device data protection.
Enabling Click to Do in Windows 11 Settings
If Click to Do is not active by default on your Copilot+ device, you can toggle and configure it through System Settings:
- Open Settings by pressing Windows + I.
- Navigate to System in the left sidebar.
- Select Click to Do (or locate it within AI components & Recall).
- Toggle the switch next to Click to Do to On.
- Configure trigger options, choosing whether to activate the overlay via keyboard shortcut, mouse hover, or the dedicated Copilot key.
Triggering the Click to Do Overlay
You can invoke Click to Do at any moment while working inside any application:
- Press the keyboard shortcut Windows Key + Q (or press the dedicated Copilot Key on supported keyboards).
- The screen slightly dims, and an interactive intelligence layer appears over your active workspace.
- Hover your pointer over visual elements on your screen. Click to Do dynamically recognizes text blocks, photos, web addresses, and tables.
- Click on any recognized item to reveal contextual action buttons directly adjacent to your cursor.
Contextual Actions by Content Type
The table below summarizes the automated actions surfaced by Click to Do depending on what you select on screen:
| Recognized Content | Available Actions | Associated Tool or Outcome |
|---|---|---|
| Text Block / Document | Summarise, Rewrite, Copy Text | Sends text to Copilot for condensation or rewrites tone instantly. |
| Image / Graphic | Erase Object, Remove Background | Opens the asset in Microsoft Photos or Paint with automatic AI masking. |
| Visual Subject / Product | Visual Search with Bing | Searches the web for identical products, landmarks, or visual matches. |
| Email Address | Send Email | Launches the default mail client with the address populated in the To field. |
| Web URL / Domain | Open Web Link | Opens the web address directly in Microsoft Edge without manual copying. |
| Tabular Data | Export to Excel | Extracts unselectable image tables directly into structured spreadsheet cells. |
Retouching Images Directly from the Desktop
When reviewing graphics, PDFs, or online media, Click to Do provides instant image manipulation without launching dedicated photo editing software manually:
- Activate Click to Do using Windows Key + Q while viewing an image.
- Click the image to select it.
- Select Remove Background from the pop-up options menu. Click to Do instantly segments the primary subject and provides a transparent PNG ready to paste into presentation decks or chats.
- Alternatively, select Erase Object to brush over watermarks, photobombers, or visual blemishes using the on-device generative fill pipeline.
Privacy and Data Isolation Controls
Because screen analysis involves sensitive workplace or personal information, Windows 11 enforces strict security boundaries around Click to Do:
- 100% On-Device Execution: All optical character recognition (OCR) and computer vision segmentation run exclusively on your physical NPU. No screen frames or text snippets are transmitted to external cloud servers.
- Exclusion Lists: In Settings > Privacy & security, you can add sensitive websites (such as banking portals) or proprietary enterprise software to an automatic exclusion list, preventing Click to Do from capturing or analyzing those windows.
- InPrivate Browsing Respect: When browsing in Microsoft Edge InPrivate or Chrome Incognito mode, Click to Do automatically disables screen capture overlays to prevent accidental data logging.