Windows 11’s Click to Do: Revolutionizing Desktop AI with On-Device Power

Windows 11’s latest wave of innovation has ushered in a new era of desktop computing—one where AI enhances everyday workflows in real time, right at your fingertips. At the center of this progress is “Click to Do,” a feature designed exclusively for Copilot+ PCs. This is not just another productivity widget tucked away in your Start menu. Instead, it’s a context-aware AI assistant working natively on your device, transforming any desktop activity into actionable shortcuts in just a click or two. As Microsoft’s ambitions for on-device AI crystallize, “Click to Do” stands out as a compelling showcase of what’s now possible, what’s next, and just how far the line between user and assistant has blurred.

What is “Click to Do” on Windows 11?​

At its core, “Click to Do” is an AI-driven contextual actions engine. Unlike “Microsoft To Do,” the familiar cloud-based task manager, “Click to Do” operates wholly on-device—an important distinction both for performance and privacy. The engine works by taking a snapshot of your current desktop and running advanced AI models directly on your hardware. These models identify both text and visual elements visible on your screen, presenting you with relevant quick-action options that morph depending on what you’ve selected.
There’s no uploading or processing of your desktop content in Microsoft’s cloud as this initial recognition is performed locally, leveraging the power of the Neural Processing Unit (NPU) found on Copilot+ PCs. Only when a user explicitly decides to invoke a cloud service—for example, searching the web or asking a web-based Copilot question—does any data leave your device.
This on-device exclusivity requires modern hardware. Specifically, “Click to Do” is only available on PCs equipped with an NPU, as seen in newer Qualcomm Snapdragon X Series chips and select AMD and Intel AI-enabled processors. If you’re using a traditional desktop or a laptop without this dedicated AI silicon, this feature—and its privacy-by-design premise—remains out of reach.

Setting Up and Launching “Click to Do”​

Microsoft has designed “Click to Do” to be as unobtrusive or as always-available as you want. Activation is simple:

• Enable/Disable: Go to Settings > Privacy & Security > Click to Do, and use the toggle switch to turn it on or off.
• Access Methods: The feature can be accessed as a standalone app (from Start, Snipping Tool’s “Click to Do” button, or keyboard shortcuts) or inside the Recall app—which itself is part of Microsoft’s new AI-powered timeline and search experience.

The first time you launch the app, it quickly captures a snapshot of your desktop. At the top of the screen, a “Click to Do” toolbar appears, featuring a search box, microphone option, and direct links to open chosen snippets in the Recall app. Right away, the interface makes it easy to begin working with whatever’s on your screen.

Core Capabilities: Contextual AI Actions at Your Fingertips​

Where “Click to Do” truly shines is in its dynamic, context-sensitive actions menu. Depending on whether you select text or an object (such as an image), the available options shift to best fit your intent.

For Text Selections​

Selecting a block of at least ten words and right-clicking reveals a suite of AI-powered actions:

• Copy/Open with: Instantly copy text or open it in relevant apps.
• Search the web: Seamlessly perform an online search, launching your browser (data is sent to the cloud only here).
• Ask Copilot: Summon the web-based Copilot for richer contextual answers.
• Summarize/Rewriting: Have the AI quickly produce a summary or a rewritten version of the selected text.
• Bulleted Lists: Automatically format the text into handy bullet points.

Depending on your installed apps and Microsoft account sign-in status, you may see additional options like:

• Practice in Reading Coach (for educational flows)
• Read with Immersive Reader (for accessibility improvements)

Microsoft is intent on broadening this AI action palette. Features on the roadmap (and rolling out in stages) include:

• Scheduling meetings and sending messages in Teams
• Integrating with M365 Copilot apps for drafting, converting, or organizing content in Office applications

For Object (Image) Selections​

When an image or object is picked and right-clicked, AI actions expand to:

• Copy, Save as, Share: Standard quick-share and save tools.
• Open with / Visual Search with Bing: Leverage both local and web-powered tools for discovery and editing.
• Ask Copilot: Engage Copilot with the visual context.
• Background/Visual Editing: Options to blur backgrounds, erase objects (with Photos app), or remove backgrounds altogether (with Paint).

The breadth of these context-sensitive actions is notable and constantly evolving. Because these capabilities may rely on features from external apps (like Paint or Photos), it’s wise to regularly check for app updates via the Microsoft Store. Without the latest versions, some options may be unavailable or grayed out.

Deep Integration with Windows Recall​

Much like “Click to Do,” Recall is an AI-centric Windows tool built to help you resurface anything you’ve seen or done on your PC. Within this timeline, you can open a past snapshot and enable “Click to Do” right from that context. The same AI-driven context menu appears—empowering you to interact with old content as if it were live, from searching the web to rewriting or summarizing emails and notes from weeks ago.

Privacy: What Actually Stays On-Device?​

By default, everything “Click to Do” interprets happens locally on your PC. Microsoft emphasizes this distinction to set users at ease: your desktop images, text, or objects aren’t analyzed in the cloud unless you explicitly choose to use a connected service (search, Copilot, etc).
Of course, the moment you invoke a web search or cloud-powered Copilot, your selected snippet or image is securely transmitted for processing. For privacy-conscious users, this distinction is vital and represents a strong step forward compared to older AI assistant models, which frequently depended on cloud analysis for even the most trivial contextual cues.
Since “Click to Do” is first rolling out with the May 2025 security update for Windows 11, only supported hardware is eligible. The privacy promise holds as long as users remain aware of which actions stay local and which require cloud transmission—a distinction that, depending on future UI updates, might benefit from even clearer labeling.

Strengths: A Quantum Leap for On-Device AI​

“Click to Do” marks a profound leap in making AI feel seamless and ever-present without being intrusive or dependent on external servers. Here’s a breakdown of the most significant benefits:

Ultra-Responsive, On-Device AI​

By using the NPU, AI tasks that previously required slow round-trips to the cloud now execute almost instantly. Summaries, rewrites, visual identifications, and context-aware suggestions appear in a heartbeat. This speed, combined with near-absolute privacy, raises the bar for what users should expect from modern personal computing.

Tight Integration and Context Awareness​

Rather than being siloed in its own window or app, “Click to Do” acts like a layer over your whole desktop. Anywhere you select and right-click, context-sensitive actions bloom. This unified approach means the utility of AI extends to every corner of your workflow—email, browser, documents, or even educational and creative tools.

Adaptive and Upgradable Action Sets​

As Microsoft and third-party developers innovate, the list of quick actions will grow. Because these actions are delivered (and updated) via app integrations and Windows updates, “Click to Do” isn’t static. Expect support for new file types, richer Office and Teams integrations, and possibly even next-generation creative tools as the ecosystem builds.

Minimal Learning Curve​

The “right-click for quick actions” paradigm is so familiar that discovering what “Click to Do” can offer requires almost no learning. No esoteric keyboard shortcuts or multi-layered settings menus—just select, right-click, and see what’s now possible.

A Boon for Accessibility and Efficiency​

Features like “Read with Immersive Reader” or “Practice in Reading Coach” underscore how on-device AI can democratize productivity and learning, making advanced reading and content manipulation more accessible to users with varying abilities.

Risks and Concerns: Caveats Worth Considering​

Even as “Click to Do” sets a bold new direction, several limitations and unanswered questions remain.

Hardware Exclusivity and Fragmentation​

The NPU requirement means only a small subset of users—those with the latest Copilot+ PCs—can even experience “Click to Do.” Those with otherwise powerful hardware but lacking an NPU are cut off. Early adopter fragmentation could be frustrating, potentially dampening developer enthusiasm for rapid expansion.

Privacy and Data Boundaries​

While Microsoft is clear that content recognition happens on-device, the transparency about when data leaves your PC (for instance, subtle differences between “Ask Copilot” or “Open with” versus “Summarize” or “Rewrite”) may not be sufficient for everyone. As AI models and integrations expand, maintaining the user’s ability to know (and control) where their data travels will become even more critical.

Action Availability and Software Dependencies​

Some AI actions, especially image manipulation or Office file handling, require you to have the latest versions of associated Microsoft apps installed. If an app update is delayed or withheld (due to region or IT policies), those features may remain inaccessible, leading to a fragmented experience.

Potential for Contextual Misreads​

AI, even with on-device benefits, remains susceptible to context errors. If the local model misidentifies content, the suggested actions may be irrelevant or even misleading. Early reviews suggest the selection and object recognition is quite accurate, but no AI system is perfect—especially as more complex documents or non-standard formats become commonplace.

Dependence on Microsoft Account​

The majority of advanced actions require Microsoft account authentication. For users who prefer local or domain-only accounts—or who are wary of tying all productivity to a single ecosystem—this could present an unwelcome barrier.

The Road Ahead: Microsoft’s Vision and the Copilot+ Ecosystem​

“Click to Do” is being launched amidst Microsoft’s larger push for Copilot+ PCs and the promise of hardware-accelerated AI experiences. Its debut alongside Recall as part of the Windows 11 May 2025 update is no accident. Microsoft is betting that the PC of the future is an “AI PC,” where routine tasks are continually streamlined, context is king, and user privacy is baked in at the silicon level.
Rumors (and some confirmed roadmaps) suggest that both third-party developers and Microsoft’s own Office and productivity suites will quickly act on the momentum. Soon, drafting documents, scheduling meetings, or even complex creative workflows may all be accessible as right-click, context-driven AI actions—no more app-hopping, no more copy-pasting between tools.
However, for now, the experience is limited to those early adopters with the newest eligible hardware, a reality that may persist until AI NPUs are the standard rather than the exception. Windows’ storied history is littered with innovations that began as luxury, then trickled down—increasingly, that seems to be Microsoft’s plan for AI.

Conclusion: Is “Click to Do” the Future of Desktop Productivity?​

In many ways, “Click to Do” represents the realization of long-standing dreams in desktop computing: contextual, privacy-first AI that feels like an extension of your intentions rather than a black box in a faraway server farm. Its strengths—on-device privacy, blistering speed, deep context, limitless upgradability—are hard to overstate. The simplicity of invoking powerful actions via right-click lowers the barrier to discovery and repeated use.
Yet, early limitations around hardware requirements, evolving privacy guardrails, and dependency on updated software environments leave room for the ecosystem to mature. As more users migrate to Copilot+ PCs, these stumbling blocks may fade—but in the meantime, they remain nontrivial.
For those with eligible hardware, though, “Click to Do” is not just a glimpse of computing’s future—it’s here, it works, and it has the potential to permanently raise expectations for how AI belongs on the desktop. Just as the mouse and Start menu once redefined user interaction, so too might “Click to Do” usher in an era where right-clicking is all you need to bridge human intent with machine intelligence.
Windows 11 users and IT professionals alike should watch closely: the age of practical, supercharged on-device AI has arrived, and “Click to Do” is just the beginning.

---

Source: Windows Central I can show you how to use "Click to Do" in Windows 11 to perform quick tasks with AI on a Copilot+ PC