Tired of digging through endless folders for the right image? In this episode of The Ultimate SharePoint Content AI Series, we show you how to use SharePoint’s Image Tagger to automatically tag your images—making them fully searchable in SharePoint, Copilot, and Microsoft Purview.
📸 We’ll walk through enabling Image Tagging, how it works behind the scenes, and how it integrates with Autofill Columns, Organizational Asset Libraries, and even raw image metadata.
🔍 This isn’t just about search—it’s about making your content work harder with AI. And best of all? It’s almost free.
➡️ Perfect for marketing, communications, and any team managing a large image library in Microsoft 365.
📌 This is part of our Ultimate SharePoint Content AI series — check out the full playlist to go even deeper.
Watch my 100+ courses on Pluralsight
Video Summary
- Every organization has an image library—but without proper tagging, it’s a nightmare to search through, especially when Copilot tries to help in apps like PowerPoint.
- SharePoint’s Image Tagger (part of Content AI) automatically tags images with relevant keywords, making them instantly searchable by both humans and Copilot.
- It only takes three clicks to enable, and it costs just $0.01 per 100 images—basically peanuts compared to the time you’d spend tagging manually.
- It supports common formats like JPEG and PNG, as well as RAW camera files, and even extracts existing metadata from the image itself.
- You can combine it with Autofill columns to generate accessible image descriptions, perfect for promoting inclusivity and compliance.
For more information, read the transcript blog below, or watch the video above!
Transcript
If there is one thing that every single company has, it’s an image library. All the stock pictures you receive, all your different logos, and all the elements you use in PowerPoints are likely stored in a document library in SharePoint. The other truth is that, unfortunately, it probably looks like this one. Uh, it has the name of wherever you downloaded, an ugly ID, and finding stuff inside is impossible.
Well, of course, putting those images in SharePoint is a good idea. You’re also supposed to tag what they’re about because if you don’t tag them properly, it can be impossible to search to find the right image. You might spend money to buy images that you already own just because you’re not able to find the image that you wanted. And even worse, that $30 per user per month that you spend on Copilot when people want to add images in PowerPoint, it might return bad results, getting people mad, not only at it, but also at Copilot.
Imagine if you could make all your images searchable based on automatic tags. Wouldn’t that be awesome? With the image tagger part of SharePoint content AI, you can now extract information and text from images. And not only will it make it searchable, like you see over here where I search for group and I have a result, but also, Microsoft now allows you to connect Copilot to your organizational asset library. And don’t just listen to me, look at the documentation. One of the things that can go wrong is insufficient tagging. If you have tags under images, Copilot can understand them and provide you with better results for your users, whether they are in the new create app or in PowerPoint.
And did I mention it takes only three clicks to do? And it’s almost free. I want to call it free. There is a cost, but it’s almost free. Welcome back, everyone, to the ultimate SharePoint content AI series. In this episode, we will talk about image tagging, and I’m joined by my good friend, Microsoft MVP RD MCT, and I cannot say all of the titles, Gokan Ozcifci. I don’t think there is enough space on the screen for all of your titles. Gokan, I feel like we have talked together in this series so much about extraction, and there are different ways to do it. All of them with different complication levels. But the difference for this one is that it is made specifically for images. One more slide.
And for those of you who know Gokan, you know how much he hates slides. He almost fell asleep with just the five slides I had. He’s more of a demo man. But one more slide and we’re going to the demo, Gokan. I promise. Okay. The only thing you need to have before you’re able to do what I’m doing in the demo right now is that you must have pay as you go billing set up. Image tagging must be turned on either for all the sites in the tenant or for the SharePoint site that you are using. And if ever you have no idea what to do that, episode three in this series is dedicated to how to configure pay as you go billing, how to properly activate content AI services, and you’ll have a link to that episode in the description below. Okay, I promise no more slides. Let’s go to the demo.
So, because I really want to show you how easy it is, I will go to a brand new SharePoint site here. Again, I’ll just refresh it to show you I didn’t do anything. I’ll open up a new document library. Let’s go to the blank library here and call it stock images as an example. Nothing else. Let’s just click on create. Awesome. We’ll give it a few seconds over here. And after that, we need to enable the image tagger on this library. It’s not something that is done by default. And it’s really important to enable it as soon as possible. I’ll tell you why in a few seconds, but enable it as soon as you know you’re going to want to use it. You simply go to the tweet docs over here. Then, under automate, you click on configure image tagger. You do one click over here. Do you want to enable the image tagger on this library? Enable.
And then that’s it. One, two, three clicks indeed. No more one-click because this is not a click. Then the second, and then the third. I told you it’s perfect. Three clicks. So you see now it added this column over here called image tags.
What I’m going to do now is, you see, I have a bunch of shutter stock images here. Some of them didn’t download. So I’ll just skip those. I’ll upload all of the rest. We are on conference Wi-Fi right now. So, you know what? And you know how stock images are when you download them from a professional website. They’re all like 40 megs each because why not. They’re like in super high definition. So, while this is done, let me go back to the slides a bit. Gokan, you don’t mind? Let me go back to the slides a bit and cover some of the details.
Gokan, I know you are a photographer. I love taking pictures, but I have no idea about all those formats you’re showing me on screen now. Basically, a lot of formats are supported. I think that is the short version. And what I really love about this is that you have the consumer formats, if I can call them that. So the JPEG, PNG, TIFF, the high efficiency ones, things like that. But it also supports raw camera formats. So if you have a dedicated photographer inside your organization, which many people do in marketing and communications, you can upload the raw image files, and then those can also get tagged.
A few things to know. It will take a Microsoft minute for everything to be done. And for those of you who watch the series, you know what a Microsoft minute is. It is anywhere between 5 minutes and 24 hours. Gokan, I think you sacrificed the SharePoint site before we started recording. This week is more towards the five minutes. So hopefully it will be a bit faster. But don’t worry, especially if you upload like a thousand of them at a time, it might take up to 24 hours, but it’s not like you have to watch it do it right. You can go do other stuff. That’s the beauty of it. You can then edit the tags. So you’re not locked in. It will give you a starting point, and after that, you can just go in and edit the tags. Add the things that you know are missing. Humans are still needed. We are not completely replaced yet.
Also, something that is amazing is the information from the image metadata. So we know a lot of photographers add metadata inside the image that will be copied over to that new column. So if somebody already did that effort to do it, it is not wasted. Now, Gokan, I know we have this problem with many other things. We have it with OCR, we have it with taxonomy tagging models? With AI models, only images added after the feature is enabled are processed. I wish they had changed that. I wish they enabled us, like an autofill column, to just click on it and process it. But this is why I told you a few minutes ago, as soon as you know you want to use this feature in the library, enable it because you don’t want to upload a ton of images and then regret it after, have to remove them, read them, and all of that. It’s not fun. And the image tags column is a managed metadata column. So, that is quite cool.
I talked about autofill columns, but you can also combine them. What if I can combine it with an image tagger to do a summary of the image or to describe it in a way that is accessible to people? So hopefully I did enough slides. Do you think it’s uploaded yet? I think so. Yeah.
Okay, it uploaded. Almost all of them uploaded. I do not have any image tags yet. So it might take a bit of time. The thing that is also a bit, well, I like it a bit less about image tags is that you don’t have the processing status. One of them wasn’t uploaded. It’s okay. We have enough demos here. Probably a network issue or something like that. But you don’t have a processing status column which I don’t know about you but that’s a general thing with content AI like if we talk about taxonomy tagging image tagging AI models like even autofill we miss that piece of showing us what’s going on like what do I wait like is it doing something autofill has that column with in progress that’s a good step and I know Microsoft is also working to have that status bar. We need it. Like now I have no idea what’s going on, right? Yeah. As an admin, I hate not knowing, you know. Let me do a quick refresh, cuz I don’t like to see the red at the bottom. Okay, you didn’t do it yet. Meanwhile, let’s try to combine a bit with autofill. And for those of you who are new to the series, this is the first episode. You see, we did a full episode on autofill. It’s an amazing functionality. So, if you like what I’m going to do really quickly here, make sure you watch that episode because you’re going to love it.
So, I will call it accessibility. I probably spelled Okay, Chrome didn’t tell me I spelled this wrong. Here, let me go to autofill, and I’ll give it the prompt describe this image for someone who cannot see it. Let’s click on text. No, the suggested prompt here was not the best one. Let’s click on the result over here.
This document contains a calendar for December and January with various appointments. Wow. Let me pick another one here. Let’s just do a test. I love how detailed. Yeah, I really want to see that image. Like is—I think it was the first one, right? It was the first one over here. Let me save it. Yeah. And then save over here. You see, I was able to talk enough to make it work, but let’s show you the image that it had. It said it had a calendar with certain appointments on it and everything. We’re just going to let the conference Wi-Fi do its thing and open the image. If not, what is the ID? One-five here. I’ll try to find it based on the description. Is it this one? One-two-five. Oh my gosh. Is it this? Let me go here. December and January. Yeah, December and January. Oh my god. So, it was able to go there with various appointments. That description was amazing. That is amazing.
Anyway, go back here. Let me do a quick refresh for good luck. You see, this one got processed first. I’m not sure if my kicking off the autofill on it had anything to do with it or not. But again, it’s normal. It can take 5 minutes to 24 hours. But you see, it gave me the image tags here: date, software, screen, office, working hours, communication, and event. And it will take a bit of time. But here, if I do “events,” I’m not sure it’s going to work yet. It did. It’s in the index. So, I can search for events. I can search for the screen. I want an image with a screen, and now I’m able to find this. And again, not only me—Copilot, your agents as well—because you now have this metadata over here.
Go, I want to do a test. I want to go back here to accessibility. I want to see if I do a test with autofill columns, will it kick off the image tagger? Do you think that’s going to work? Let’s try it. Go on the second one. Let’s go on the second one. Let’s do a test. I wonder if that forces it to process it, and it uses the same in the back. If it does, I just found a brand new way to demo it faster. So, this—okay. So, wow. W JPEG. Okay. Last modified… Yeah, it was just luck. That’s the fun about doing live demos. You get to test things. You know what works, what doesn’t. We are not necessarily going to wait for everything to work. You see here it worked for one of them, and the rest of them are going to appear in the future. Again, it can take up to 24 hours, especially since I uploaded a few of them at the same time.
At the beginning of the video, I told you it’s almost free, and I did not lie to you. How much does it cost? Goen—peanuts or 0.00001 or something like that. There we go. 0.01. So that is a tenth of a cent in USD. So how much is the work to do this by hand? It’s not worth not enabling this. And each processed image counts as one transaction. So it’s easy. Think of it as—you need a hundred images, a thousand images to cost you $1. Enable it. There’s no reason for you not to enable it. What we’re going to do is we’re going to add the image here, the screenshot, and you’re going to see how long ago the image was modified. So, this will give you all a realistic good time on how much it took for everything to be here.
But Gokan, anything else you want to add about the image tagger? As a conclusion, as Vlad said, it would be crazy not to use it for all the advantages you get compared to the price. It’s like you have to use it. That’s a conclusion I would say for that video.
Awesome. Well, for all of you, if you have enjoyed this video, make sure you like the video, subscribe to the channel, and if you want more ways to use AI in SharePoint, make sure you check out the rest of the videos in the series. You’ll have the next one appear on your screen right about now. And again, we have about 17 episodes in this series. Some of them might not be out yet. Make sure you hit that subscribe button to get notified as soon as they come out. And see you in the next one.
