In recent months, we have seen how big technology companies are pushing hard Artificial Intelligence in their everyday products. Microsoft is no exception: Windows 11 and its apps Native apps are being filled with smart features that, beyond the hype, seek to solve everyday tasks with less effort.
One of the most striking new features lands in the Photos application: a automatic categorization powered by AI that promises bringing order to chaotic librariesThis feature, still in testing for Insider Program users, classifies images locally on computers Copilot+ PC, with useful results from the first minute.
What is Microsoft Photos Auto-Categorization and how does it work?
Microsoft has started testing a feature in Windows 11 that AI built into the Photos app which automatically analyzes the content of your images and groups them into predefined thematic folders. The company calls it Auto-Categorization or Automatic Categorization, and it's available starting with version 2025.11090.25001.0 of the app (or higher) in Insider channels.
For now, the system works with a set of fixed categories that cover the most common desktop categories. According to Microsoft, the AI is capable of detecting each type regardless of the language of the text appearing in the image, and even distinguishes between manuscript and printed documentsThe categories available today are:
- Screenshots
- Receipts
- Identity documents
- Handwritten notes or memos
The key here is the robustness of the model: AI identifies visual patterns typical of a receipt, a screenshot, or an ID card, without the text needing to be in a specific language. For example, a passport written in Hungarian, Japanese, or Arabic will still be correctly labeled as an ID document.
This organization is done automatically and without you having to manually move files to folders, an alternative if you are looking for other options are the programs to organize photos. Photos in Windows 11 Scans the library and places each item in its place, facilitating searches and reducing the “noise” of the giant libraries we accumulate with There.
It's worth noting that, for now, the scope is focused on documents and screenshots, not people, pets, or locations. No face recognition or custom categories At this stage, this is something that many users will miss for their more personal photos.
User interface, search and control
Once the feature is activated, you will see the new folders inside the left sidebar, under the drop-down menu. "Categories"From there, you can go directly to "Receipts," "Screenshots," "ID Documents," or "Notes," without searching blindly by date or system folders.
Furthermore, the search bar Photos integrates with these categories to quickly locate any item. If you manage hundreds or thousands of images, this quick access makes all the difference: locating a ticket for a return or a screenshot with key data becomes a matter of seconds.
AI is not a closed box: if you detect an error, you can manually modify the category of an image. This user fix helps keep the library clean and, in the long run, contributes to improved overall performance as the feature evolves.
Microsoft encourages suggestions and corrections via the Feedback Hub, which aims for continuous improvement. Although the feature is in the testing phase, the focus is clear: save time, reduce clutter, and make it easier to navigate your library without changing your manual albums nor your existing folder structure.
If you came from other tools, you will notice differences. Google Photos bet on dynamic tags and semantic search, while Microsoft's proposal focuses on fixed categories oriented towards desktop work. It is also partly reminiscent of Windows Recall due to its visual component, but here the function is limited exclusively to organize images within the Photos app, without extending to other areas of the system.
Requirements, availability and privacy: what you need to use it
Self-Categorization is progressively deployed in all Windows Insider channels and, for now, it is exclusive to Copilot+ PC devices. This means that you need a computer certified with hardware specific to AI, in particular a NPU capable of exceeding 40 TOPS (trillions of operations per second) to run models locally and efficiently.
The update comes via Microsoft Store and requires the Microsoft Photos app 2025.11090.25001.0 or later. You can wait for it to install automatically or force the update from the store, though don't be surprised if it takes a few days to appear: the rollout is in batches and may vary between devices.
As for availability by equipment type, the feature is active in portable, tablets and some mini-PC within the Copilot+ ecosystem. Some point to a future expansion into desktops as new generations of CPUs/NPUs arrive, especially on platforms Intel and AMD, but today this is a rumor without official confirmation.
Regarding privacy, Microsoft hasn't specified in detail whether the analysis is always performed on the device or if there is a cloud component. Since the feature is limited to Copilot+ PCs with NPUs, it's reasonable to assume a local processing by design, which reassures those who handle sensitive images such as personal documents and receiptsHowever, it remains to be seen exactly whether there will be granular on/off switches and how user control will be managed.
When it comes to control and productivity, the value is already there: the Photos app reduces repetitive tasks and lets you find what's important instantly. For work teams and personal use Where “digital paperwork” accumulates, this layer of AI brings order without asking for additional time or changing your flow.
Super Resolution: AI image enhancement built into Photos
Along with automatic categorization, Microsoft has expanded the deployment of Super Resolution on Copilot+ devices with compatible Snapdragon, AMD, and Intel processors. This feature uses AI models to sharpen and upscale images with stunning quality, reaching multiply by eight the resolution in some cases.
The integration is done within the Photos app itself, so you don't need external tools to "rescue" old captures or low-quality photos. The system can offer several levels of improvement, similar to what solutions like waifu2x and other popular utilities do, but here with the advantage of being embedded in the Windows user's workflow.
To start it, the app will request the download of a model package specific to your computer; once installed, processing is done locally using the NPU. This is good news in terms of performance and privacy, with tight response times and without the need for a permanent internet connection.
By combining Auto-Categorization and Super Resolution, Windows 11 Photos covers two key fronts: organization and visual qualityYou can automatically categorize and, whenever you're interested, enhance images on the fly that you want to keep, print, or share to make them look better.
It's important to remember that both new features are being released gradually. If you don't see them yet, check the Microsoft Store, update the Photos app, and see if your device meets the Copilot+ requirements. Access often appears first for a subset of Insiders and expands gradually.
Advantages, current limitations, and how it compares to Google and Apple
In everyday life, Auto-Categorization stands out because it attacks what is most annoying on the desktop: receipts, screenshots, IDs and notes that get mixed in with the rest. Automatically separating them is a relief when you're preparing an expense report, looking for a purchase receipt, or need a scanned document without digging through thousands of files.
Now, the approach is different from that of Google Photos and Apple Photos. Microsoft is betting on fixed categories and not by open tags or deep semantic search. This makes it very practical for desktop workflows, but it may fall short for those who want to classify by people, pets, or places, which are the quintessential "emotional" photos.
Another difference is the scope: this function lives within the Photos app and it doesn't extend to the entire system, as might be expected from a general-purpose "Recall." It's a more limited approach, but also more manageable and easier for the average user to understand.
Among the current limits, there are two clear requests from the community: enable custom categories and expand the recognized types (for example, specific invoices, course materials, or product photos). Microsoft hasn't announced anything firm, but it would be a logical step in future iterations.
Furthermore, the multilingual model and the distinction between manuscript and print mark a before and after in productivity. That a Hungarian passport is classified well Even if you don't understand the text, it's exactly what you'd expect from a well-trained AI for documents.
Passionate writer about the world of bytes and technology in general. I love sharing my knowledge through writing, and that's what I'll do on this blog, show you all the most interesting things about gadgets, software, hardware, tech trends, and more. My goal is to help you navigate the digital world in a simple and entertaining way.