Free Copilot Chat Lands For All, Exchange Audit Logging Changes & More: The Practical 365 Podcast S4 E34

Join Steve Goodman and Paul Robichaux to discuss Microsoft's latest announcements, including the surprise move to make Copilot Chat free for all Microsoft 365 users, significant changes coming to Exchange Online audit logging, and the new Outlook's arrival via Windows Update. Plus, get their takes on Microsoft's major AI reorganization with the new CoreAI division, and what all these changes mean for IT pros as we start 2025.

Want to stay up to date on all things Practical 365?
 Follow us on Twitter, Facebook, and Linkedin to stay up to date on all things Microsoft!

Transcript

Copy

0:09

Season 4 Episode 34 of the Practical 365 Podcast with me, Steve Goodman and Paul Rovershaw. Hi Paul, how are you? Hello from Copenhagen Steve I'm fine thanks how are you? I am good. I think it's colder where I am than it where you are. Without the witch or Fanta. I mean, that's entirely possible, but it's dark for both of us, which is rare. Usually one or the other of us will be in darkness. I know, I know. So, well I was going say happy new year before we into dark metaphors. If you know Copenhagen then you'll probably be envious of Paul, except the fact it's the evening and he's still working, but out his window he tells me is the lovely canals where on a summer's day you would go out to get an ice cream or perhaps go out for a lovely steak or whatever food you would like. Yes, in fact, the neighborhood where I am, I'm about a six or seven minute walk from the district known as New Haven, which is an iconic street that has all the Dutch looking colored houses because it was originally designed by Dutch architects as was this canal system in Copenhagen. That makes so much sense. You know, I've never thought to look up or ask. It is a lovely place that you're in though. It's your second home, I think. Yeah. Maybe, maybe a three minute walk from here. There's a delightful ice cream place and then almost right across the street from it is a craft beer bar on a boat. So yes, in the summer, this is a perfect place to be right now. It's almost on the corner. Yeah. It's called isobar. You probably have. I'm not, I'm not sure it's a landmark exactly, but I think it's pretty well known. They're ice cream. Boy, it's good. I wish I had some right now. In fact, Yes. And if you're in sunnier place in the world, then go enjoy some ice cream. Actually, before we start the podcast, my daughter had some ice cream. That must have been what was making me subconsciously think about it. Yes, I was like, yes, it's just over there, Don't call me darling. Anyway, I digress. The first podcast of the new year, is and we've been saying a few times on the podcast. Well, we predict all sorts of things, because I suppose if you predict everything, you're bound to get something right. But The cost of Microsoft serving Copilot may well be reduced over time. It certainly seems that Microsoft are feeling very generous or they're feeling very eager to get everybody's hands on Copilot in Microsoft 365 because Copilot chat. is now part of Microsoft 365. I hate to say free because you've still got to buy it. maybe is a better way to put it. What do think of this port? You could look at it as what was the enterprise Bing chat. It had a lot of similarities with copilot chat for access to your Microsoft 365 data. Is this as big as it sounds? You know, it's medium big, I think I would say. we've had, just from a technical standpoint, we've had the ability to do chat-based reasoning that's grounded in tenant data. We've had that for a while. Part of what's new about this is it doesn't require you to have a copilot license, but it does allow you to use agents, which is a huge focus for Microsoft right now. Maybe... not coincidentally, you can use agents that are provisioned under a pay as you go plan so that the more your users use those agents, the more consumption that drives and the more money Microsoft gets. So I'm not saying that having this tool be available at no extra cost is related to that, but I'm not saying it's not. Well, it sounds very much like AI companies and Microsoft and Emeta and so on would like this to be the year of agents. And certainly, there's a big ask there, isn't it? Because if an agent is going to be particularly useful, what bar is it going to have to meet to make it valuable enough? But it seems like a really good addition. by Microsoft into getting people to be able to consume in a pay as you go way agents. Will that be successful? You I don't know. You know, we didn't see that with syntax, for example, when it moved to pay as you go model, at least certainly not overnight. Yeah, you know, it was interesting because if you look at syntax, just kind of as a broad technology portfolio, I think, it's a pretty good example of Microsoft trying to take some things that were adjacent to each other and put them together in a unified brand and then sort of forgetting what the unified brand was for. So they've been very successful with the branding by an intra. They've been very, very successful with the brand behind purview and defender. and even Microsoft 365. But with Syntex, you had this kind of weird mix of some things that were storage related and some things that were content AI related, although they weren't really called that at the time. Well, because we started with what became Viva topics and SharePoint syntaxes, two things that were launched that were adjacent together as separate products. And then there's like a lot of things, right? They go find their own home. content classification and tagging and things that seem almost. Not naughty now in a way, because. Easy understanding of documents is almost part of most modern generative AI models as well. Especially with the ability to use tools and things like that, venture and calling. But the point being that Microsoft started that as a license, didn't they? For SharePoint syntax. Then they moved it to a pay as you go model. But the on-ramp to use it was quite high. There's an on-ramp to use these agents, is build an agent that begins life as something you build inside Copilot Chat, then get other people to use it in the organization. It's got a model for consumption. It's an opportunity in a way, but it's quite a limited one, right? Yeah, it is. But you know, I like the idea that Microsoft is going to enable people to be, gosh, I don't want to call them citizen developers, but I don't know what, you know, you call an agent builder, but they're democratizing use of this inside the enterprise. And I think that's really important because they can't, Microsoft themselves can't build every capability and every, you know, agent behavior that people might want. And the world we're in today is still one where You have to be pretty knowledgeable to be able to do your own model training and build your own model based solutions from scratch. It's certainly doable and it's much easier now than it was say two years ago, but it's still not to the point necessarily where somebody in the accounting department or the shipping department is able to easily do it. So this lowers the entry bar quite a bit. And I think that's good. I think we're going to see it get some uptake fairly quickly. Businesses want lower staff costs. Someone starting a startup today might say I'd like to employ half the number of employees as my competitors. So give me an advantage. But it is that isn't it? know if you could make it easier for. The the overworked person in a know our best case scenario overworked person in department. They need three people. They're never going to get three people. It's just them if they can. Automate. the most laborious parts of their job using agents, then that's great. I mean, the goal of AI companies is to be able to automate the entirety of particular jobs that are suited to do so, That doesn't mean people will choose to do that or choose, you know, but some industries will. They just feel a little bit like turkeys voting for Christmas. kind of the most pervasive use case that I think, and for whatever reason, I'm not sure that there are any AI companies who really capitalizing on this per se. But to me, the biggest and best and most general use case is using the AI to do things that are too tedious or complicated for you to easily do. And I'll give you an example of something that I did today because I was trapped in engineering meetings all day, which were really good, but also you know, fairly high velocity. I had several pictures of whiteboards of lists of stuff and I needed to turn those into a format that we can all collaborate on. Well, naturally, first thing I wanted to do was, hey, let me put this in an Excel sheet. So I dutifully opened up Copilot in OneDrive and asked it to extract the text from those images. Well, Copilot can't do that yet. So then I tried Chat DPT. I have the free subscription. don't pay open AI because I think they are very close behind Google in the, say don't be evil, but then let's be evil anyway, sweepstakes. So I just don't want to give them any money right now. Despite the fact that Microsoft is giving them money. by extension, when I pay Microsoft anyway, a chat GPT couldn't do it either, but, uh, Anthropics called model did it very easily. I upload uploaded the image and said convert each. block of text here into a entry and it did a very good job. made very few mistakes. So let's leave aside the maybe five minutes it took me to decide that those other two tools couldn't do it. I was done in perhaps five minutes, whereas retyping all those things by hand would have taken me probably half an hour or more. If you multiply, you multiply that. an opportunity for IT departments to break free of not necessarily. I've said this a few times before on the podcast that sometimes there's a little bit too many restrictions on what copilot can do depending on your organization. Right. know, Microsoft will sell to a company. Um, but also that does, it makes guns. Um, but won't talk about, make a marketing document for those guns. Fair enough, okay. We've had folk on the show, on the podcast to why. But if you do want to make it do those things, or you want to give it capabilities like Paul's mentioned, that don't exist in that ecosystem, then it's a good opportunity to do so. Because GitHub Copilot does now give you the opportunity in preview to use the Claude Sonic model, right? And it works very well. You can use it to work on a code base and do some really sort of interesting things that sort of rival the open source equivalents like Klein for coding. Which means that you could build using a copilot agent that did exactly as Paul describes, know, takes that photo, uses a state of the art model that more expensive for a transaction to return that result. I use it for much more tedious stuff as well but like my like we're walking around the shop supermarket and my daughter's done a shopping list my wife has just shouted out to her bread milk blah blah blah whole long list of all the sort of things and and my wife said right let's all go off in three directions to get all the food for Christmas. And I'm looking at this and going, okay, this makes no sense to me because it's not in any order. So I took the, I a photo of my daughter's phone screen with this list that she'd taken a photo of. I uploaded it to the Claude and asked it to write it as a shopping list based on the, based on the normal layout of a UK supermarket, which it did, based on the aisle layout. Yeah. it got it completely right actually for a handwritten list. And I thought that was pretty cool, right? You know, that's making sense of things that are sort of intangible that syntax couldn't do very easily. You know, that's super unstructured data. And that's almost similar to sort of your use case in a way where a state-of-the-art model can do these things like magic. Yeah, I like your point about being able to have a greater degree of customization. And I wonder if at some point the commercial situation between Microsoft and OpenAI is going to lead Microsoft to decide that they would like to give their customers more platform choice. And I don't know, you know, it's impossible to say how that'll work out. That's like so far above our level as, you know, 365 admins, it is sort of comical, but you can see that there might be some business value in letting Customers choose a tier of model that has more or fewer capabilities depending on what it is they want to do. I mean, you can, you've got access to other vendors models in Azure alongside OpenAI. AWS, Bedrock is it, they have Claude Sonnet access as well. And I think it's, I think GitHub Copilot is offering Claude Sonnet via AWS if I remember correctly. So quite the sort of. curve curve curve it your way of going around to get that access. So yeah, I I wouldn't they you know what Microsoft need to Microsoft providing office and copilot. Is it is it a good thing? I know the launch announcement says it's using GPT 4 under the hood. That's a smaller model right than chat GPT 4. It's not using anything state of the art for reasoning. What if you want? What if you want that? What if Microsoft aren't aren't the right folk to provide it? What if you're you're looking at cheaper access to other state of the art models from some of the big Chinese vendors like DeepMind? where you can get access for a fraction of the cost. For a large organization, especially an international organization that's not massively concerned about data centers on US soil, or can host those models themselves, then why not? It would be a great use of being able to give access to these kind of functions. But just through... the Microsoft 365 copilot chat. Because the least interesting part of this is building a front end to it, isn't it? Yeah, pretty much. Although I think a close second in terms of least interesting is there's a fairly new copilot feature that will summarize a document for you when you open it. Now I've seen this for a while on insider builds, but it looks like it is getting out in the wild. And I'm curious what you think about it. I'm. I think it's fantastic. I mean, theoretically, don't see the harm of it. My kids, my daughter would say to me that this is environmentally unfriendly. And I would disagree. I'd say it's not, cumulative, this has probably got some impact. But you know, but. The kind of things that kids are getting told at school, know, about AIs environmental impact is somewhat out of date. And. You know, I don't think it costs Microsoft a lot to use something like GPT-4 mini, which is said to be. I think that's probably true. And I guess I would just say generally as someone who has put four kids so far with one more nearly done through the US education system. If you think they're being misled about environmental impact of AI, boy, wait till you find out about some of the other stuff they've been misinformed about. But and I'm not joking when I say that. You're right that the aggregate impact is pretty big, but I guess when I think about this from a standpoint of me as a user, I've got to admit, I'm kind of struggling to see the value. So first of all, many of the documents that are going to get opened in my tenant or in my work tenant are going to be documents I've created. So there is absolutely zero value to burning even one CPU cycle to summarize it because I already know what's in it because I wrote it. Yeah. So leaving that aside, because I understand not everybody is on the output side of the equation. Many people are, you know, they spend their jobs consuming documents that other people have created. I guess I just have to wonder whether this is a feature that actually genuinely delivers value. If every time you open a document, you have to pause for some amount of time. Well, pause is the wrong word because it opens in what appears to be a separate thread and it doesn't block you from accessing the document. I guess I'm not completely convinced that it offers enough value to have it be on by default. And as far as I can tell, there's no way for you to turn it off from a tenant standpoint. Individual users can disable it. But I don't know that an admin can. it? How do you effectively manage or measure usage if it's being manipulated? Is it being manipulated? it a waste of time or is it something that on balance does something which is always hard in adoption? Reminding users that the feature exists For some people, know, sometimes that might be like, oh, I forgot I could do that. For other people, might like, can see that. know, to my way of thinking, and I have to admit, I haven't tested summarization in Copilot thoroughly in a while, but when I did test it before, it led to, you know, that infamous Hacker News article. That was fun. What I found is if I ask Outlook to summarize something using Copilot, if I ask it to summarize the same message 10 times, I would get variable results. Now, this is not unexpected because, you if I ask my grandfather to tell me a story 10 times, he's not going to tell exactly the same story all 10 times. are going to be times when he focuses on one aspect of the story or another more, or he gives me a longer version or a shorter version. So I get that it's not necessarily deterministic and I won't always get exactly the same summary. But if I'm talking about document summarization and I think, okay, I'm going to open a 25 page document. How good a job is that summary going to do accurately reflecting? enough of the document so that the summary is useful to me? And the answer is I don't know because I haven't tested it. I mean, technically, if they're going to do this and not make it enable by default, then the most efficient way and cost effective way doing this is usually through batch inference anyway. why don't they summarize these upon change if it's always going to be done. really, whether you ask for a document summary or whether it's automatically showing it, it's always there in the metadata. And then bring out common tags that the organization uses. Because Viva Topics is never really going to sell ever, it? But if you actually go to the trouble of doing a summary of a document, then you could also, the same call could ask against all of the organization's SharePoint metadata, all the taxonomy that exists. It would cost Microsoft nothing more if they're going to do this. Because obviously if someone saves a document, they will open it again in the future, right? So schedule for batch inference and save. People used to SharePoint doing things as scheduled tasks. It's the cost-effective way of doing those kind of transactions. I'm just looking for a positive way of doing it that actually would give a benefit to our organization, right? it. It's cool. Right. I mean, the idea that I can have a machine give me a useful, accurate, factual summary of a document written in natural language, that would have been science fiction 10 years ago. Right. Nobody could do that until fairly recently. So the fact that I have that on my, you know, my desktop, that's pretty dang cool. It happens to be a kind of cool. doesn't personally benefit me. but I can still appreciate the accomplishment of it, right? I mean, it's a nifty feature and I know there are people that this will be useful for. Yes, I'm just thinking it could do, you know, it's not exactly state of the art. It's in the, for a lot of people, it's in the category of party tricks, right? It's obviously on one step above, ask a co-pilot to describe you based on what it knows as a superhero or something like that. Really lame and pointless stuff that people waste their workdays doing. And that's going from someone that hasn't tried this themselves. The voice of experience. above doing stuff like that, but it's a party trick isn't it? to some people summarisation is a party trick, to others it's really useful. Does it need to be on by default? I don't know. I don't know whether Microsoft are actually doing it on document load, if they're doing it on save. they're queuing them for batch summarization. So it's super efficient, has to be done once. Then that has no impact at all, you can test that because the feature will work if you open a document that's an email attachment, for example. So when you open that, there's not really a SharePoint site to host it. And I think all the document summarization today is being done on demand. That's not to say that they couldn't have a scheduled job that does batch inference. That might be more efficient in some cases. But you know this really well. I just don't understand why anyone wouldn't. Well, I can tell you why. Because if you think of all the people that have got dusty SharePoint data that they only retain because they don't want to do classification or cleanup, you're going to end up burning a lot of cycles to summarize documents that no one will ever open again. No, mean on safe, it. If it's a document that's been saved, queue it for that. Or if it's opened and then summarized, have a property that's used for this that is internal, only used by copilot, whatever it is, stored in shard as static metadata. But whenever I do any sort of captioning, classification, anything I do, I always try and keep. so that I never have to do it more than once because it costs money, right? If I'm writing software that's going to go and do any sort of inference, batch inference is cheaper. And if the answer is going to be roughly the same every single time, unless it changes, then it would seem wildly inefficient to, for example, send out an email attachment or a link to a SharePoint document to a hundred thousand people. They all open the document and then it summarizes it a hundred thousand times. That would be, you know, We've been talking about this feature for five minutes and come across that as a problem. If they wrote this and didn't think this through, it must do this. If it didn't, well, perhaps that's a bigger problem to do with Microsoft's dev teams and their focus on AI. Maybe they focus more on getting it right. you wanted to get this capability in market and see if it's valuable and useful, one way to do it would be you do it on document load. And then if people like it and it seems like it's useful, then you come up with a more elegant way to do it at better scale. This isn't a test pluff when people are paying for this. sure, but that doesn't mean it's not a test platform. I I know you've used windows over the years, so let's not pretend like everything that Microsoft ships is fully baked and not just picking on Microsoft. The same is true for Google and Apple and you know, to put Gmail beta in it and Microsoft would say, oh, this isn't a beta. getting what you paid for, right? Yes, I know I'm just being silly. does half baked features are not top of my priority list, right? Why make something wasteful? That's, you know, it's like building a tap but not a faucet, water faucet and not being able to switch it off. Because, well, let's see if people want to switch it on first. We could have really interesting conversation about the death of optimization in computer science as a whole, because as we move to a world where compute and RAM and network connectivity are so ubiquitous and so cheap, man, there are a lot of people who are writing software and using tools that would have been unacceptable 10 or 15 years ago because they were so resource intensive. So that's probably a topic for another time though, because it'll lead us to talking about the metaverse. Well, it is a good segue into power platform. Which I low code no code environments are designed to be that way, right? They're not efficient in any sense because. They're another level above right? But one of the problems that everybody has with that kind of stuff is tenant migrations and I've done. I've done quite a few where it's relied upon. manual export and re-import of applications relying upon users to move some of their own. Come February public preview of being able to migrate these tenant to using PowerShell command that's coming to the service. I think that's a fantastic thing. If you were beginning planning an on-demand migration project and this was one of the things on your list that actually you didn't have a full solution for, it might be worth looking at whether The public preview is something that you can join and perhaps use for that migration, you know, alongside the proper migration tools. Yeah, I like that they're doing that because one thing a lot of people who had not used a power platform or if they're not steeped in the ecosystem, then they may not be familiar with the concept of what an environment is and power platform and data verse. So you can almost think of this, an environment is sort of the counterpart object to a tenant in some ways, although you can have multiple environments that are tenant. They're usually used to say, okay, I have a... an environment for developing software and then I've got a staging environment and then I've got my production environment. But because you can have multiple environments, the ability to move stuff from one environment to another, even if they're in different tenants is pretty significant. It's going to enable quite a few professional app, you know, app lifecycle management scenarios that we haven't had up until now. Yeah, think it's a really good thing. must be a long-waited engineering effort as well. think some things were reliant on Microsoft to be able to do and this is certainly one of those. So yeah, I think that's absolutely fantastic. Thankfully, I'm not doing tentative migrations these days, but if I was, then... Yeah, it could be something I'd definitely want to get on the preview for. I'll look forward to hearing your report once it gets little closer to release time. Now there's news that Microsoft are reorganizing their dev teams around AI services. And you work with a lot of developers, don't you? Is this? they don't like being reorganized. No. However, it seems perhaps better than what Mark Zuckerberg was recently saying on one of podcasts that he joins and talks on about how, know, CEOs love to say about how AI is going to replace people. But having someone say it's going to replace thousands of my employees would be so demoralizing compared to, it's, it's worse than that because he also said they were laying off the bottom 5 % of their performers. And so what that means is now each one of those people, when they go to look for another job and they're asked, why did you leave Facebook? Why I left Facebook? Because I got laid off. you're one of the losers that they got rid of. That's just mean. There was no reason for Zuck to say that as part of the announcement of the layoffs. But that's always been happening, hasn't it? You know, that happens at Microsoft and other companies, Six Sigma and things like that have been around for a long, time. And people still are only up from having ex meta, ex exoogler, exoogler, whatever they call it, or ex Microsoft on the CVs. You know, I know people who worked at Microsoft 20 years ago and they still go, it's still almost the second thing that they say, you know, I work at so and so I'm at ex Microsoft. So I don't... That's not the worst thing. The fact is you were employed there in the first place. You know a lot of stuff. The barn 5 % is still in many people's opinion, probably quite higher than the top 10 % of a small software development company. Yeah, true or otherwise, right? So it's a much after doing a big reorganization, so not super popular compared to not doing a reorganization, but I suppose it's not that thing is it? They do need to reorganize towards AI, but. I wonder whether this means there'll be more unfinished. Things in the closet bits of software that. You know, that's always a possibility. It's super interesting to me if you look at what got put together. They took the existing development division and they added to it the AI platform division, which is what owned a bunch of bits and bobs of AI. And they stuck it together with some teams from the office of the CTO, which I thought was interesting. So the ones that they list in the blog post are... their AI, AI, agentic runtime and engineering thrive. Man, that sounds good, right? I want my engineers to thrive. Who doesn't? So, so I'm wondering, you know, if this is the precursor, very much like any other kind of organizational game, whether it's organizing the cups and saucers in your cabinets or playing Tetris. When you move pieces around that creates new spaces and new configurations that you can use to put more stuff. So is this an indicator that we're going to see other real ornaments? Maybe, maybe not. mean, who knows? It's criminal criminology to even ask about it. But it is interesting that they picked these specific pieces of the organization to plug together in the new order. I mean, it could be a very, what's the word, tactful way or corporate presented way of the same thing that Mark Zuckerberg is saying, which is preparing to be able to do a lot more coding through the use of AI agents, right? That's obviously going to happen because I mean, who wouldn't use GitHub Codespaces to do a bit of that? It's soon, if not already, a dev's core skill. It'd be able to use AI in a way that doesn't take more time than just doing it by hand. So could be that. wish the reorganizees the very best of luck in their new homes. I mean the thing about reorganisation, having been through these myself, right, you sometimes you don't win out of them, sometimes you do, but they always offer opportunity, right, to step up of some kind. you know if you weren't, if you were in one of these roles where you weren't doing AI and now you're going to have the opportunity to be part of that, then that's, you know, that's always going to be a win whenever there's some new emerging technology. And some of the wins in science with AI have been combining two separate things where AI, well, the knowledge from one and the knowledge from another hasn't really been combined. And that could be where Microsoft see these wins, where they've got the minds of people who are really not focused on anything to do with AI, extremely smart people combined with folks who are quite adept at using it. And then they're working together and that's been some, you know, because it's been brilliant. Perhaps they'll set AI agents to crush all bugs in Windows 11 or something like that. Exactly, you know, just churning away at all the stuff that people don't have time to deal with, right? So I think positive things. Yeah, I mean, I think, I think it's always important to make sure that your organizational structure reflects what you're trying to do. And so given what Microsoft has said about the structure and also what they're, you know, what we think they're trying to accomplish. And I think these are probably a good move. And if they fix windows 11 bugs automatically with agents, and that's just gravy on the ice cream. Wait. we were talking, we complained about the about Copilot doing a document summary and talking Windows 11, we forgot to mention Windows Recall, which is constantly doing this, you know, on the PC side. So in the scale of things. we've given recall such a hard time already. Probably don't need to recap that, but yeah, you're right. It's a difference of scale, right? If it's warming up the CPU in my local machine to summarize documents on my local machine, that seems like that's different in some important ways than taking a random document and asking the service to summarize it for me. You're right that in the end, either way, you're being presented with a summer you didn't ask for and can't stop. So, we've got quite a lot on the show today. Yeah, you know what? One other thing I want to bring up because I miss this the first time because I was asleep. Microsoft had a bit of an MFA outage on Monday of this week. It was Monday the 13th, which is not usually a bad luck day, but it was interesting because it regionalized to Western Europe. So it actually happened while I was in the air asleep on my way to Copenhagen, which is why I missed it. But I'm curious, did it affect you at all or was it too far west or sorry too far east of you? I wasn't affected by it, then again, I'm logged in and using it all the time, I didn't get any people who use my tenant phoning me up in a panic, which they usually would do. Often the MFA attitudes don't affect... you can have very large organisations. I remember when there was a big one, affected a very large UK telecommunications company, one of the IT folks who I knew there, and they said that the impact was not as massive as you would think, even with hundreds of thousands of people. But to that one person, it does affect, it's massive. Yes. And I think now that Microsoft is putting more emphasis on broader deployment of MFA, I know they've invested a huge amount in resilience and stability and service quality for MFA and the results have been very good. Right. That's probably the part of the service that has the fewest problems. If you look at each individual component and look at, you know, Exchange Online versus MFA, for example, the MFA guys end up looking pretty good. I mean one MFA thing that did change for me today actually is SMS messages began, at least in the UK, moving to WhatsApp. So yeah, earlier in the week SMS and now the user gets a message saying if they obviously got WhatsApp, check WhatsApp for your one time code, one time codes and now send there. It's a good thing. I say this because I changed my phone. And when you change your phone, it's a pain, isn't it? Setting everything up again. So I the SMS notifications on as I recently changed my phone. I haven't signed everything back up. So I log into the Quest 10 and I got the notification. It's like, oh, OK. So it's over in WhatsApp. And it works, right? It works fine. And of course, if the users had these before, then they're not going to get a WhatsApp out the blue. They get that final message that's in the same thread as the existing messages with the one-time codes, the verification codes as before. it's, even to a lazy loadout like myself, hadn't said authenticator again, I found it wasn't any hassle and... then suddenly it became a lot easier because now I just see the WhatsApp notification and my phone lets me copy and paste it in. So if you do have users in that situation, then this will happen. They may ask if this is genuine, as they probably should. And we'll get a link to that if you weren't aware of that rolling out. it's a good change because so much has been written about SMS security and multifactorial identification over the years that some organisations who have had to keep it unwillingly will be glad to see the back of it. I know I will. If I never get another SMS authentication verification message, that would make me very happy because it would mean that all the apps that I use have either switched over to TOTP in an app or pass keys. So we should probably have some time in an upcoming episode dedicated to talk about pass keys because there's a lot going on there in intra and in the broader industry, Windows and Mac OS and so on. So we'll have to note that down for a future topic, but I think now we're about out of time. Yes, so yeah thank you for listening to the show today, first one of 2025. We've got some interesting news for the next show but I will leave that for now. Check out the blog for the show where I'll double check that I can reveal all about our new co-host that will be joining us on the next show. and needles. yes, and we've already got some really good guests lined up as well. So yeah, happy new you better make sure you subscribe to our podcast at practical365.com or wherever you get your podcasts from so you don't miss out. Yes, and thanks for listening and I'll catch you in a couple of weeks time. Yep, happy doodier everybody.