Microsoft's Copilot Is Changing Fast (Here’s What Actually Matters)
Everybody, welcome back to Make Others Successful, a podcast where we share insights, stories, and strategies about how to build a better workplace. We got a fun one here. We're talking about AI as is everybody else in the world, but it's for good reason. We do have some changing thoughts on Copilot and the way Microsoft is is moving, and I think it might be interesting to to talk through. So if you haven't met me yet, my name is Mitch.
Mitch:I'm happy to be joined with Mike and Matt here. The usuals if if you've been around, but let's talk about Copilot. So this world is moving a million miles an hour, and it feels like we have even sometimes struggled to keep up, especially to the point where we are able to formulate our thoughts around these tools and then articulate them to somebody. And so it's just this constant game of something just happened before this meeting that we learned about coming from Microsoft that is changing the way we think about it. We can't talk about it yet, but it's it it triggers things in our brain that says, oh, maybe we should integrate that into how we talk about these tools.
Mitch:And so it's it's a moving, crazy, fast world. So we're gonna try to take a snapshot of, right now, our thoughts about what Microsoft has been doing and where they're going and what we're excited about and what things that we think they should pursue or consider. So let's talk about Copilot's we're calling it new direction, like, our thinking about how we think about Copilot in the realm of AI and tools and LLMs and and things have have changed. Does someone wanna take a crack at that new thought process?
Matt:I would not call it necessarily new, but I would call it more formal, like they've actually done it. Copilot has never been an LLM. When people compared it to ChatGPT or to, you know, any of the the models and they're like, oh, it's just a different version of this or whatever. Copilot doesn't have its own LLM. Copilot uses other LLMs to improve their products and suites and deal with their data that they have.
Matt:Can can
Mitch:I say it's it's fair? People have probably thought in the past, should I buy Claude or should I buy Copilot or should I buy ChatGPT? Stuff like that.
Matt:Yeah. And and that there's still you can still have that conversation that's not going away. But the reality is Copilot is just integrating more models. It's always used other models. It's never been its own thing.
Matt:And now rather than just being ChatGPT, which anybody who's watched this for any Copilot for any any length of time, they know ChatGPT was the basis for all of the
Mitch:Huge partnership.
Matt:Huge partnership. And and it was always, like, the reason to use ChatGPT over Copilot is you get the latest model faster. Right? Because it's it is still just ChatGPT on the back end from an LLM perspective. And now, you've got Claude being incorporated in.
Matt:You've got ChatGPT. You've got like, they're starting to expand that those partnerships and those involvements to basically get kind of the best of all worlds. So
Mike:is fair to say then that when you're trying to ask and answer that question, should I buy should I pay for ChatGPT? Should I pay for Claude? Should I pay for Copilot? You're what you're saying, Matt, is, well, if I have Microsoft three sixty five and I pay for Copilot, I have access to Claude?
Matt:Yes. Today. Obviously, Claude's been around a while. It took them nine months, a year to really get that integration to really work well. Right?
Matt:So, yes, today, yes, there are still reasons you might want to buy Claude separately. These are separate tools. They have separate integrations. But when you go, what I care about is the model. The model.
Matt:Right?
Mike:So the the reasoning power.
Matt:Yeah. The reasoning. You have access to both. Right? You have access to all of those things.
Matt:The real differentiation now is data. Claude, by default, doesn't have access to your m three sixty five data, although you can easily add it and grant access to it, and the integration with the other tools. Right? People many times are taking stuff out of email, out of the other piece of data, putting it into Claude, generating something, taking it back out of that, putting it somewhere else. Like, it is a a thing.
Matt:They've created lots of integrations to, like, PowerPoint and Excel, etcetera, but those are all add ins that you have to install and manage. Copilot is all included. It comes integrated into all those tools.
Mike:So is it worth having the conversation about DLP and governance when it comes to allowing a tool like Claude to access my Microsoft three sixty five content? It's probably one
Matt:of the biggest areas to think about and how important it is.
Mitch:Just to find someone who doesn't know DLP.
Matt:Yep. Yep. I'm getting
Mitch:there. Sorry.
Matt:So what what Mike's talking about is the security around the data and the usage and how that all works, both of the large language model and the AI functionality and the data that you're leveraging with that large language model is really important. In Microsoft world, DLP, data loss prevention policies, stuff that tries to stop people from sending data to somewhere else, compliance policies, retention policies, all of these things are all built into the Microsoft suite and automatically apply to everything that you're doing within the Microsoft three sixty five suite, including Copilot things. And so it creates a single pain, if you will, to do all that, to manage it, to govern it. Right? Now, I'm not gonna sit here and say it's perfect.
Matt:Microsoft has lots of challenges where this one works a little bit different than this area over here. It's not a perfect kind of situation, but it's it is still one platform, one vendor that you're dealing with. When you think about the other landscape, organizations who two years ago, were all bought in and shot GPT, and that's where they went, and that's where they they had a bunch of data. And now they're like, well, we don't like that one. The new one's clawed.
Matt:Let's go to that one. Now they've got a clawed environment with a bunch of data there, And now there's gonna be a new one in six months that people are trying to go into. Well, Gemini. Right? And we've now invested in Gemini.
Matt:We've got seven different ones that we're investing in. Each of those is a different surface that needs to get managed. Each one of those is a different surface that needs to get protected and governed, and it just creates more and more and more. However, the Microsoft tools are never gonna be the newest, latest, greatest. They're just not.
Matt:Like, they're not creating the model. They're not creating that unique thought in the marketplace these other companies are. And so it's always gonna be behind. So if you're looking for the bleeding technology, if you're looking for the bleeding is an interesting word in this to your point earlier, Mitch, about the pace of things. When people think about that, they think I think they have a lot of conception about what that is.
Matt:But in the AI world, you bleed for a little bit of time and then all a sudden it becomes everybody's doing it.
Mitch:Yeah. You're late.
Matt:Because the pace of advancement is so aggressive. But
Mitch:I think part of that comes from when something new feels more magic than the last thing. All of a sudden, everybody wants that thing. And so it people are trying to get there, and then it's this constant game of how do we be magic again.
Matt:I would I would equate it to like the Joneses. Right? And there's this is another big thing, is there's like a bunch of different levels here. Some people just want to use the next thing because it's the next thing. And they want to feel cool and they want to feel like they're using the cool thing that everybody's using that is out in the news and people are talking about and
Mitch:All of your LinkedIn and their hacky
Matt:speaker that I listen to on a podcast once every week is, like, raving about how it's it's saved them so much time. And, well, we've gotta use it. We've gotta use it. And then there's people who are, I'm just trying to get stuff done. I have to like, my bosses are telling me I need to, you know, use these tools to cut my work down in half.
Matt:Right? And I need to figure out how to do that, and I need access to these things to figure it out. Right? And then there's people who are like, I don't want to take any risk. It needs to be a 100% rock solid.
Matt:It needs to be well integrated in my work, and I just that's what I need. Right? So there's lots of different levels there. But, yeah, I mean, it's it's a interesting situation.
Mitch:Yeah. If I had to sort of timeline out the last year or talk about Microsoft's moves in it, it feels like they hitched a ride on OpenAI. That's a loose way of saying, invested tons of money into that and and and integrating into their tools. It never quite felt like magic in the Microsoft tools. It always felt like it could sort of interact a little bit, but never really felt like it could do things for you.
Mitch:All of a sudden, Claude, Anthropic is making moves and they're they're going ahead. And Microsoft says, oh, shoot. We need to get on that bandwagon too. Let's start integrating with that tool. And I think Matt has talked about before, this is going to continue to be a race to the finish, and there is no finish where there's gonna be something new that Microsoft is gonna wanna hitch onto and and integrate into the tools.
Mitch:And I think, generally speaking, it's a good direction. If I were to go back a year, I would have assumed Microsoft was trying to build their own LLM. They may be. I don't know. As a in getting out of the OpenAI relationship, but it seems like they're taking a little bit more open approach where they're integrating rather than trying to lock something down.
Matt:The smartest move for them. The reality is the investment in in that they need to make is all about integrating with their tools and making it a wonderful seamless experience for their the people who are using those tools. Yeah. And I they don't they don't have time to build their own LLM and deal with all of that thing. This is my opinion.
Mitch:That's Apple has kinda figured that out too as well.
Matt:Like, don't Why not just use what all these other people are investing in and use it for for my purpose? Yeah. I also think they've gotten kind of lucky because both OpenAI and Anthropic are not the same as some of the other models that are seen as a little bit more dangerous in the in the marketplace from a privacy compliance. They both have very strong stances on those things.
Mitch:Yeah. So let's talk about kind of how this has translated into the tools and the things that that we're seeing. I'll say the first thing that I saw was in my Copilot app. I saw a drop down for choosing your model and it's normally on auto or something. And I clicked that and all of sudden I saw Sonnet, the Claude model.
Mitch:I said, oh, that's interesting. And that was kind of my first hint at they're starting to integrate this into other tools as well as when they started to talk about agent mode in the apps. We found out that's done through the the Claude platform as well. But do you wanna talk a little bit about let's talk about agent mode, and then we're gonna talk about co work next. That that was exciting too.
Mitch:But what does what does agent mode being introduced into apps mean?
Matt:I mean, it's this is all about the the new things that are coming out or have been coming out. And agent mode is really a transition, a transformation from I can interrogate and ask questions to I can do stuff. What we've had in the past is largely I can ask questions about a thing, I can ask it to create content, it's generative, it's I can create a new thing, but we've struggled using Copilot to do anything meaningful in regards to changing things, changing an Excel file, changing a PowerPoint, changing anything. It's been a challenge.
Mitch:It would be like generate take one crack at generating five PowerPoint slides and then be like, there you go. Yeah. That's what it was before.
Matt:And none of it could be templatized. None of it could be something that somebody wants to use in the end. Right? Yeah. We used it a lot for ideation.
Matt:Like, I would ask it to create something, it would give me an idea for what the slides could be, and then I would go spend a bunch of time making my own stuff. We're finally gonna get what, you know, some people have been using for probably six months to a year now with some of the cloud models, with being able to just say, hey, create a PowerPoint, and it's gonna create a good enough PowerPoint. Like, it's gonna be good enough for in probably 90% of the scenarios for what we want. And being able to ask it, I've been using it to ask stuff in Excel. Right?
Matt:I have a big table of data. Create me a pivot table that does whatever. Analyze this thing and look for this. Right? Still is challenging.
Matt:I think you need to learn how to use it. But definitely, it is it's changed the way we think about things.
Mitch:Mhmm.
Matt:So
Mitch:Yeah. I I feel like one of my first experiences with it was trying to sort out some some time logs that we had in regards to a project and trying to figure out how to attribute them exactly. And there was long descriptions in each entry, and I was like, I don't really wanna, like, think about this. Can you help me categorize things in these three different ways? And Austin's like, sure.
Mitch:I can do that for you. And it creates new rows, highlights ones, and creates a legend of how to read it and gives me all the the actual formulas right in the Excel, and it was, like, done in thirty seconds. And I was like, I did not have to pros I didn't have to use my brain to process that that busy work that it could do for me. And so that is where I feel it. It was like magic.
Mitch:It just did it for me and did not feel like I needed to do its job. Matt, did you
Mike:mean to imply when with your last statement that the Sonnet LLM or that model is what has made that work better, that agent
Matt:work better this is another thing that's changed a lot is the namings. It's not agent mode anymore. It's editing with Copilot or whatever. Mhmm. They've been trying to do this agent mode thing, like, call it agent mode in the past and all the stuff before.
Matt:I don't know what they were using it with before. Yeah. Today, I know it's using Claude. Like, that is that's what Yeah.
Mike:So I was I'm trying to connect.
Matt:I don't I don't know whether or not they tried to do it something else and then decided it wasn't good enough. Like, I don't know. I can tell you that all of the current agent mode functionality or editing functionality and the co work functionality, which we're going talk about in a minute, is all predicated on the Claude models.
Mike:So it's a theory that the introduction of the Claude models has enabled that to be better?
Matt:I mean, I think there's without a doubt that's that is a k. True statement that is better.
Mitch:It's what has enabled it to happen. Yes. It's it's almost not even a comparison of better
Matt:or worse. But that's but that's not unique in the marketplace. ChatGPT can't do that. Right. Like, there is that is one of the differentiators is that Claude is able to do work.
Mike:So if you're paying paying for Copilot and you're using Claude models or Anthropic models, what's the incentive for Anthropic to continue to allow Microsoft to leverage that platform?
Matt:I mean, they're getting paid a lot of money. Okay. Like, I like and and it's yet to be seen whether or not Microsoft's gonna put limits on like, they already have limits. The the licensing model and like, there's a lot of questions about how this really rolls out in regards to money. And just like everything else, it's changing all the time.
Matt:If you've watched any of the news information, Microsoft rolled back on having included Copilot for Yeah. Large organizations. Right? They are renaming m three sixty five Copilot to Copilot Premium. And they're, like, they're doing all sorts of things and changing that world, and it's changing just as fast as a technology.
Matt:Like, I would not be surprised if they said, much like in Researcher, that when you have a regular Copilot license, you get a certain amount of usage, but after that, you can't get any more. Why? Because they're paying they're paying those Anthropic isn't doing this out of the for charity.
Mike:Right.
Matt:They're getting paid every time somebody makes a request and it's doing work
Mitch:for And their
Mike:the reason and the reason I'm asking that is because I'm trying to figure out, like, what's the right mindset or approach for an organization who's trying to figure out how to adopt the right tools for their team members to be able to use to actually get get work done. And if they're like, yeah, we're gonna do Copilot, and then they invest a bunch of time and effort, and then Anthropic decides to pull the rug, like Do mean But it doesn't sound like that would happen.
Matt:I mean, I don't Anything's possible. But your question is, I think, the root of your question is a really good one, which is that you need to have a strategy for doing this. Organizations that allow their users to go choose whatever model they want, any way they want, need to be prepared for the consequences of that. Users or organizations who are super restrictive in, you know, you can only use this model and that's all we're doing should also be prepared for the consequences. All of these choices about all of this stuff in this changing landscape that's moving a 100 miles an hour has consequences.
Matt:Right? We've talked to customers who have made the choice to that we're only we're we're focused on Claude. Right? Mhmm. And we said, well, just so you know, that's like that that's just the current model.
Matt:And, you know, there could be a future one that's better. And the response is, well, that's okay. We're going to stick with this one. But I think if you ask that same customer six months ago, they would have said the same thing about Chatuchi BT. We're going to choose Chatuchi BT because it's so much better than Copilot, and that's the one we're going to use.
Matt:And then they would have been stuck in this world where they have multiple, multiple, multiple of these things they're maintaining, which just increases this risk exposure and increases the cost. What are you trying to get out of using AI? What is your business goal? What are you trying to support your users doing? And then solve those problems.
Matt:That's what you need to be focused on. And if that requires you to have ChatuchPet and Claude and you know, Gemini and Great. That's awesome. Do it. I love it.
Mitch:I feel like some of the evolution of our thinking has of should I buy Claude? Should I buy Copilot? Not being exactly apples to apples helps us think about maybe Copilot is a pretty safe, easy button for a lot of these solutions. Because they get the models that are from the other companies, and it's all integrated with your your existing tenant and your information, and it's safe. And it might not be bleeding edge, like it takes a little bit to catch up, a couple months behind when Claude does something, but it's it's a lot easier than managing three different ones
Matt:or so.
Mitch:Especially if you are trying to, like, dip your toe into this world and you're worried about things like security and
Matt:So another customer of ours that we were talking to has kind of maybe come up with or they're thinking about it in, you know, included Copilot, included AI, premium Copilot maybe is another level. And then there's like, I'm going to use whatever these other models are. So another big thing that people want to think about, Claude, if you actually care about your data, it is incredibly expensive. Like, because people think about these things and they go, well, we're going to buy a pro license or a max license. Those things have very little protection for your data.
Matt:Like, they have protection. You can go in and manually turn off on your client and say, I don't want you to use my thing to train your model. Yay. But that data, you're still sending to them, and there's very little legal repercussions if that if something happens with that data. Or it's not an enterprise thing.
Matt:It's an individual commercial license. Right? Well, then you step to team, which that's not crazy. It's a little bit more expensive, but it's not crazy. Then you step to enterprise.
Matt:Now you can't even get it without 50 licenses. And each of those licenses are very expensive. Like, it's but then that is where you get HIPAA compliance. That's where you get, like, all the things that businesses that are doing anything that is sensitive in nature, that's where they need to be and it's incredibly expensive. Mhmm.
Mitch:I want to talk about price and things like that in in a couple minutes. Want to come back to that. Before we get too far down that road, I want to talk about co work Copilot co work. Yeah. So it was early this year, early twenty six, Claude announced this or released this product called Cowork, which is a very cool concept.
Mitch:So how I would view it is instead of something being an asynchronous, like, linear process where you are interacting with the tool regularly, a yeah. A chat. It is able to say, I see all of these different things that could be done. Let me go do those things for you, and you can just leave it there, and it will go act on your behalf for you. And point like, go click around your different tools and create an email and download new updates and things like that all for you with just a simple prompt.
Mitch:So that that concept is super strong. I have a few use cases that I'll talk about in a minute, but that is getting folded into Copilot now as well. And it's in Frontier, I think, the moment. Yep. Mhmm.
Mitch:It's called Copilot Cowork, and the concept is the same. You can ask it to do something for you, and it will go act on your behalf. Have either of you used it, tried it, have any
Matt:I've tried a little I've tried a little bit. Like, either one, Claude or Copilot, co work?
Mitch:Copilot, co work. Specifically. Yeah.
Matt:I've I've I've messed with it a little bit. I don't really have a hot take on it. It's it's only really been available for, I don't know, two weeks, something like that. It does a lot. It takes a long time.
Matt:It's kind of a little bit I I would think about it a little bit more like Researcher, if anybody's familiar with that, where
Mike:And how it behaves.
Matt:I'm not gonna it's not a thing that I'm gonna pop in and be like, put a prompt down and in two seconds I'm gonna get a response and move on. It's the kind of thing where I'm gonna spend some time, articulate a thing, send it, and then come back in an hour and I'll have something. Right? So that's that's kind of my feedback on it, I think. I think it has a tremendous amount or could be very valuable to people because it actually allows you to do complex things.
Mitch:Have you given it a shot yet?
Mike:I have, actually, today, a little bit. And I the thing that I my initial reaction was, oh, I wish it had access to a couple more Microsoft homed data sources. Uh-huh. And it would do pretty cool things for me. Like, it made some good suggestions, but it wasn't based on any planned or structured work that I know I have.
Mike:Mhmm.
Mitch:And
Mike:I'm like, oh, well, that's great that you carved out an empty time slot for me to do that, but that's not connected to when this thing needs
Matt:to be done.
Mike:Sure. And so it didn't have access to that information.
Mitch:All the context.
Mike:Right. And so I'm like, okay, how do I get that other context into a tool that it does have access to? Sure. And so now I'm like brainstorming on that. But Yep.
Mike:It definitely has potential.
Mitch:Yeah. Yeah. I'll I'll talk through a few of my well, one of the the main scenarios I've been using it for is every night, it looks back at my my day and looks forward about a week in my calendar and says, based on all of your emails, your Teams messages, your your meeting recordings, like the transcripts and things like that. Let me look ahead and make sure that the work that you need to do is allocated in your calendar and it doesn't get eaten up with meetings. And so I've watched it.
Mitch:All of a sudden, my calendar will will get booked up for a day where it says, hey, I know you need to work on this project. Let me grab this chunk of hours here totally automatically, and I don't need to think about it. And so it's able to take all of those inputs, process, and like you said, it might need a little bit more context of, I don't need three hours for that thing, I only need thirty minutes, and or the the deadline is on this date. But generally speaking, it's been pretty good. It sees, oh, I have you This client kickoff call is on this date.
Mitch:Let me grab thirty minutes earlier that day to so that you can get ready for it. And that has been great for me. I hate managing time and calendars and like
Matt:We all know that.
Mitch:Yeah. I'm annoying in many respects in that world, but it has allowed me to not really think about, am I working on the most important thing right now? It's been able to manage those things for me. So that's cool. That has been the primary one.
Mitch:The other ones are maybe more research based where it will look back at my meetings for the past week and say, hey, I noticed this common thread where, you know, everyone has this similar problem. Maybe you can generate content based on that topic, and it could be an avenue for for helping more people with with your content and things like that. So those are a few of my use cases. I have seen similar to you where in the interface, it will pull in skills that it needs to interact with the different tools. And as far as I know right now, you can't create a skill for it to So there's I tried to get it to integrate with like our sales pipeline to ActiveCampaign or Harvest, our time tracking tool.
Mitch:I feel like it might be really helpful to integrate with at some point. But it can't right now, and so it has a limited set of skills it can do. But within that context, it's able to talk to Microsoft Graph, which is the tooling that exists in between all the Microsoft tools and allows you to to interact with them under the hood. And it will basically say, hey, I can create this event for you. I can edit this event.
Mitch:Are you cool if I do that? And you can either grant it one time permission or say, it's always okay for you to create events on my calendar, which I did. And that's per prompt, per, like, conversation. You can give different instances, different permissions. So I thought that was really cool.
Mitch:And then that just reminded me of another instance of me using it, which is if I'm on a sales call or a new lead comes in and says, hey, can you send me information about this thing? It's easy to lose track of all those different threads. And so I've asked it, hey, I need to close loop on all these things. Look at my transcripts. Help me figure out what I should do.
Mitch:And it'll say, oh, I see you need to email John an overview of these things. Here's a draft email for you. Are you cool if I send it? And that concept is like it could be extremely helpful for me. But because it's new, it has some rough edges.
Mitch:So I noticed it didn't, like, pull in the the to address. Like, it didn't fill in who it was gonna email it to. It would allow me to fill it in, but that's not particularly helpful. And then at the bottom of the emails right now, says, sent from Copilot Cowork. And I'm like, I can't do that with external.
Mitch:I I'm not gonna be the can I turn that off? And it I went down and ran a hole figuring trying to figure out if I could do that. But, anyway
Mike:That seems very dangerous, actually. Like, I like the that instills a great deal of fear in me and like, yeah, sure, you can always do this on my behalf and then having to apologize a week later for the garbage that you sent.
Mitch:Yep. Yeah. Yeah. I've been very particular about what I give act I do not give it access to send anything to anybody. Yeah.
Mitch:Only do things like for me and my calendar regularly.
Mike:Yeah. I wonder if that's a like, you talk about all of the governance and the policies in the DLP, like, hey, Microsoft, that might be a policy you want IT teams to be able to like turn on or off, so that their people aren't inadvertently sending garbage emails to clients with Copilot.
Matt:Yeah. I mean, it's I had a similar experience. I asked it to summarize some things for a meeting later today, actually, and it went way down a weird rabbit hole. And it it I don't know what was different about it, but, like, even when I used regular Copilot, I was like, I couldn't quite get what I wanted to get out of it. But Cowork did took way, way long time to produce not at all what I wanted.
Matt:Sure. It looked way too broadly at random stuff related to the thing and included the stuff from meetings from weeks ago. And and it I probably wasn't as explicit. Like, I'm the worst. I do not like writing long prompts.
Matt:I'm like, if I have to write a prompt, might as well just do it myself. Like, a a a big one. Yeah. I would imagine if I had given it a better prompt, I probably would've gotten better results. But, yeah, it's it's interesting.
Mike:You probably ask it to write a prompt for you.
Mitch:I've asked AI to write prompts for me.
Matt:A 100%.
Mitch:Yeah. Yeah. I view a lot of that as, okay, cool. I see what it could do. It's definitely in Frontier.
Mitch:It's definitely in in preview mode. And I have noticed even even the interface has changed
Matt:Oh, yeah.
Mitch:The last week.
Matt:Oh, yeah.
Mitch:And they are for sure working on it, and and I'm excited to see what it could do. So that all to say, that is powered by the cloud models as well. So Microsoft making its moves, you know, integrating with other AI tools from different companies has sort of allowed them to color outside the lines and and and, you know, dream a little bit as far as how it integrates in its tools. So I'm excited about what's to come. I've seen some value from it already.
Mitch:Definitely still some rough edges. But let's go back to your pricing thing, your the cost of all these things. I was talking about it sometimes Copilot feels like an easy button where it's safe and it's secure and maybe not bleeding edge, but it's an okay solution. But it's $30 a month
Matt:Yeah.
Mitch:Per user. Let's talk about if if we could write a love letter to Microsoft of how we would structure these products. Yeah. What would you if you could wave a magic wand, what's the problem, first of all, and what what do we think Microsoft should do?
Mike:Yeah. So e seven's the new thing?
Matt:That's what they think the new thing is. Yeah. Yeah. E seven is the new thing that they're trying to do, but
Mike:What is
Matt:e seven? E seven is just a pricing model that kind of captures all of the things that you might want, including a lot of the Copilot and agent related compliance and, like, management features.
Mitch:Into your micro Into somebody's Microsoft seat.
Matt:Yep. So that you don't end up having as much, oh, I have to buy this and I have to buy this and I have to buy this. It's kind of an all in Package one deal? Kind of altogether. It's if you were bought into AI, you can buy an e seven license and you'll have most.
Matt:It's still not all, but most of the AI functions that you would want in all of your things.
Mitch:We have a note here. It's the first new license tier since 2017.
Matt:It is the first new license tier. In in the e Sure. Sure.
Mitch:G e I hate licensing.edu
Matt:you know, a models. From my perspective, I think Microsoft is missing out a lot. I think I'm very disappointed they pulled away the free or the included plan. I think the included license or the included functionality they had before was a great step in getting people to adopt.
Mitch:What was that?
Matt:So that was, we have some of the content on it already, but it's the ability to have basically web grounded Copilot chat in all of your Microsoft apps. So in Word, Excel, PowerPoint, you can interrogate documents and you can manage those things. And now, especially for larger for for smaller organizations, they're keeping it. For larger organizations, they're getting rid of it, which just makes it even more confusing.
Mitch:They're getting rid of it to kind of nudge people to pay for it instead.
Matt:I mean, they don't say that in their marketing. Oh, That's not what they say. But, yeah. So that's what I believe is what's really going on is they're not getting the adoption they want at that level.
Mitch:Yeah. So they're saying, you're 2,000 people. You can't use all this for free.
Matt:Yeah. They've removed that. I think that was great. I think that was an awesome move for them. I think We need to have
Mike:it, not awesome to remove it.
Matt:No, to have it. Yeah. To to have it for everyone. I think it was a great path on the adoption scale to what Mitch was talking about. Like, if I were talking about it, I would say add that back.
Matt:I would say the next level should be somebody that could just use it for meetings and for managing transcripts. Right now, today, you've got Teams Premium, which gives you summaries, which are kinda not super helpful if they're not right. Because now I have to take them and put them into Copilot and ask it to rewrite it or, like, try to do something else, try to grab the transcript and, like, I don't know,
Mitch:seem A very large majority of our AI use is around A meetings,
Matt:lot of the value that is a no brainer for people is when you're in Zoom, there is a license tier that gives you summaries and transcripts and asking transcripts questions and etcetera. And it is more like Copilot Chat with It's proven justifiable expense. Yeah. I mean, I I would be pushing every company to buy it for everyone that owns meetings or does meetings on a regular basis. 100%.
Matt:But I don't wouldn't do that for the Teams Premium license, which doesn't give you all of Mhmm. I recognize their problem with it, which is, like, it's it's maybe hard to figure out how to limit the Copilot chat experience and make one that's in between. But, like, if you made that for $15 a person or $10 a person, that's like, I think you could sell that all day long. To be clear, I'm not trying to shill Microsoft stuff. I think it would provide tremendous value to people, and at that price point, people could could maybe justify doing it.
Mitch:Because because we're stuck in this world where instead of saying everybody should have this, we say, you should understand what both sides looks like and which what's the risk and what's the benefit. Giving everybody Giving
Matt:today, I'd like this this is a great question. Today, our posture is you should probably never give someone Copilot full license without training, without a rollout plan, without understanding. Right? I think you could make a case for a limited feature set that is focused around meetings, that has a lot of history about success and accuracy. And you could make a case that says it's cheaper, and I probably don't have to do so much rollout, and I have to like, it's very targeted.
Matt:The moment you go to a Copilot plan, it's like all of the options in the admin center, even just getting that right, is a full time job. Right? And the changes that are going on with it. Let alone the fact that now you've got notebooks and you've got the Copilot pages and you've got, like, all these different features that I gotta train somebody on and try to Don't worry about any of that. I just wanted my meetings to be transcribed, and I want to be able to ask questions about it and get data about it without having to pay $30 a month.
Matt:Right? So I think that's that's a huge miss for them. I wish they would carve that out and make that a feature. And then in in regards to, like, Copilot itself, it really is a concept of making it more integrated. And we've said it before.
Matt:When they launched Copilot, it was not a good product. It was not useful. I don't think anybody should pay the money for it, really. And a lot of that had nothing to do with the model, nothing to do with the how good AI is or anything. It had to do with how well it was integrated into my daily work and provided value to me.
Matt:I think today, there's tremendous value. I think it's not quite a no brainer. I think the work they're doing this year will produce, to much your point about, yes, I could go out to Claude and maybe get better, or I could go out somewhere else and get better. I think with the work they're doing this year, I think you I would struggle to find a scenario where Copilot what Copilot's going to be able to do for you isn't plenty good enough for the average user. And you're probably going to be in a case where I would be recommending to organizations that large swaths of their population should just have it.
Matt:And yes, you should make the investment because you're going make that investment, but you're going get tremendous value out of that investment because of what it can finally do.
Mitch:But that still requires training. Training, compliance, governance, management,
Matt:monitoring. Like, it requires a ton of stuff. But I think previously, it still required all of that stuff, but all you could do was ask it to rewrite something at a third grade level or at twelfth grade level. Like, great. I could do that with lots of other tools and I don't have to train people, and it's like, what's the point?
Matt:Now you're going be able to do really valuable stuff, managed really well. It's gonna be great in several months.
Mitch:So right now, the two camps are too far apart free versus paying for Copilot, right, in as far as what you get, what you need to train on versus free where people can just talk with their documents. Yeah. And it's missing this, what if I could just get it for meetings and the the day to day work that doesn't include all the frills. Perhaps they're missing something Well, in the
Matt:what I think they're doing, what I think they're doing think Microsoft looked at this and I think they've heard people, because I've said it to them on many occasions, I think they're saying, I can only spend time on one thing, so I'm going make this big thing really, really good. So then Matt will say, yeah, I know I'd like to do this, and then you get off cheaper, but really just start buying this for everybody. And I think they're down a path that in six months, a year, that's gonna be the case. Like, I'm gonna be all bought in on pretty much everybody that is any sort of knowledge worker, any sort of, like, anything should have this level of license. I still think there's other pieces of population that really would need that meetings thing.
Matt:Where we've been in the past year, two years, is this thing isn't really good enough to be good for everybody. Really is for the top level knowledge worker, for the knowledge worker that is really doing complex, really Yeah. Super valuable things for your business. I was gonna say, I think
Mike:they're missing the opportunity by not having that middle tier for adoption and market share.
Matt:A 100%. Like, if they would
Mike:introduce that, they would capture so much of the market and
Matt:then A 100%.
Mike:It would be a no brainer for, you know, the the other 50% of the market that's gonna come on board
Matt:A 100%.
Mike:When the full thing becomes good or great.
Matt:And they never explained why they they made some of the changes, like, to get rid of the included version, like, are very little explanation.
Mike:Well, the fact in the fact that there's like
Matt:I mean, I only sorta know.
Mike:Small company versus big company.
Matt:Yeah.
Mike:Like, there's a fairness question in there, like like, come on.
Matt:Yeah. Yeah. Yeah. Were you gonna say?
Mitch:Oh, $30 is just too far for what you get in all the requirements around it today. If if they had a $10 bring smarts to my meetings level, I would say everybody should have it.
Mike:You know
Matt:what I'm afraid of, though? This is the thing I'm really afraid of.
Mitch:Everything becoming a la carte?
Matt:I think we might be saying the same thing. I'm worried that they're going to put co work and other things on a $60 plan.
Mitch:Yeah. Yeah. Yeah.
Matt:Now they've made their middle tier thirty, and thirty is not is still not enough for me to be like
Mitch:Yeah. Is this just a taste of what their premium plan is?
Matt:That's what was saying. It depends on if
Mike:they care more about the users or the shareholders.
Matt:I mean, it's also a position in the marketplace. Anthropic is very expensive. Yeah. Like, there's just no question. It is incredibly expensive.
Mitch:Value sell? And We're announcing here officially Copilot Ultra
Matt:coming soon. Ultra super
Mitch:yeah. We could talk about this all day. I would be interested to know if people enjoyed this, have thoughts on it. Obviously, we always say drop feedback to us at the bulb.digital/feedbacklink. I think we need to wrap it up for now, though.
Mitch:Any parting thoughts before we go?
Matt:I can't wait to talk to you, Al Baez, about what's coming next for that stuff. It's fun we know about it, yeah, that we can talk about it.
Mitch:I can't wait. Because I think Against our nature to not integrate things that we know into how we guide people.
Mike:I hope somebody on the product team for Copilot listens to this episode.
Matt:And changes what they're doing? Mhmm. Yeah. I can only hope.
Mitch:Here's your playbook, Microsoft.
Matt:Yeah.
Mitch:Right. Okay. For now, thanks, Matt. Thanks, Mike. It was good chatting about Copilot.
Mitch:We could probably talk about all of this in a brand new context in a month because things are gonna change. But when you hear this, check-in with us. See how things are going, and it's gonna be ever evolving. So, yeah, thanks for chatting for now, but we're gonna go. We'll see you next time.
Matt:Thanks. Bye, everybody.
Mitch:Hey. Thanks for tuning in to Make Others Successful. If you enjoyed this episode, we'd love to hear from you. There's a couple ways you can do that. One, we'd love for you to rate our show on your favorite podcasting app.
Mitch:And then if you have feedback or topic ideas or suggestions for future episodes, head over to bulb.digital/feedback and let us know. Your input is super valuable and we'd love to hear from you. Don't forget to subscribe and share this podcast with others who care about building a better workplace. Until next time, keep making others successful. We'll see you.
