Critical Look at Microsoft’s AI Investment  🔗 **Source:** [Critical Look at Microsoft’s AI Investment](https://share.snipd.com/episode/95aa8da7-3a9f-475c-9156-978a1342fc2f) ## Highlights - **Microsoft's Big Applied AI Play** - Microsoft created a $2.5 billion applied AI implementation unit to help enterprises configure AI and generate returns. - Connor argues the headline masks reallocations but signals large vendors are trying to solve low enterprise adoption by offering implementation teams. Transcript: Jaeden Schafer Microsoft has just committed $2.5 billion and 6,000 employees to a new AI implementation unit. Now, this is not new. They're not the only ones that have done this. They're following what OpenAI and Anthropic both have made partnerships with different organizations, Accenture and Deloitte, if I'm remembering correctly. And I mean, there's a bunch of players that have been doing this, a bunch of private equity firms that have been kind of getting in on this and kind of investing money into these sort of AI implementation units. Now, Connor has a spicy take on this that I was reading on LinkedIn. I was like, Connor, we got to talk about what you've been saying. This post has been blown up and I think it's making some people mad. But I see a lot of truth to this. So and basically, his take is that maybe what Microsoft is doing is not the right way to do this, which is, you know, shocker, especially when this is $2.5 billion that are about to be spent On this. So Connor, I'd love for you to maybe explain a little bit about what these implementation units are hoping to achieve, what you think they will achieve, what you think people should Do, and if this is all a giant $2 billion mistake. Connor Total billion dollar mistake, everybody. Yeah, I love the hot take here, man. So, by the way, I did have people from Microsoft writing to me and being like, it's not really 2.5. It's sort of like that's kind of like reallocated, but the message is the same. And even people inside Microsoft are like, yeah, you know, the headline, don't really love the headline, but essentially like the, you know, the headline is essentially, let me grab It. It's from the information originally. If you don't know the information, but it's expensive to apply to subscribe to, but it's worth it. Microsoft commits 2.5 billion to new applied AI consulting effort. And then says the new unit comes as large companies, and this is important, the new unit comes as large companies have groused that it can be tricky to configure AI software to their own Needs and generate meaningful returns. ([Time 0:00:29](https://share.snipd.com/snip/a68ccf8e-2daf-49c2-b5c3-bc784ef96d0c)) - **Energy Company SAP Rollout Flopped Because Habits Won** - Connor recounts working with a large energy company where leaders blamed the new system after an SAP rollout instead of changing how they worked. - The anecdote shows repeated corporate behavior: people resist new systems and blame the tool rather than their habits. Transcript: Connor So it reminds me, so we were working with the big, so AI Mindset like works at big companies to sort of like transform from a behavioral standpoint, right? Organizations to really transform AI adoption. And I remember working with this huge like oil and gas or energy, energy company, sorry. And we were kind of like talking about all this and doing this big senior leadership workshop, which we do. They're like, you know, three and a half hours long or something like that. And at the end, these folks were talking about when they did like their SAP enterprise transformation thing. And they're like, yeah. And remember how we had to keep on tweaking it because people just weren't really satisfied with it and all that kind of stuff. And I kind of paused the conversation. I didn't really want to get involved because it was a conversation kind of among them. But I was like, can I just point something out? Has this happened before? And they're like, yeah, it happens all the time. I'm like, this is what we think about all the time. Every time you sort of like have a new system, Jayden, a digital transformation takes place. Hey, we're now going to use Salesforce. What happens? I promise you everybody complains because people like the old system. So what do they do? They blame the system instead of how they work. ([Time 0:02:18](https://share.snipd.com/snip/8dddb44f-b804-4a3f-96bd-b75670243a34)) - **Tool Fixes Won't Drive Enterprise AI Adoption** - Connor observes AI vendors solve the 'easy' technical integration problem while enterprises actually need behavior and process change. - He warns pilots often show low usage because adoption depends on people changing routines, not just tweaking features. Transcript: Connor And I'm telling you, this is what happened. They put out a pilot of like 5,000 to a 50,000 company. And what they say is, everybody's going to use this. And then it's going to catch like wildfire. And then we're going to sell our licenses to the other 50,000. What happens? A fraction of those people use it. It's just truth. Right. And the whole point is that they are now getting feedback because they're like, well, what's happening? Why aren't people using it? They're like, oh, I don't know. I just don't like the features or something. And they complain about the thing, which I'm going to sound like a broken record here, but like, it's like complaining about the treadmill. It's like, you know what? I don't know. It's just sort of like it's downstairs or like I have to do laundry or like the features. I don't know. The problem is not the treadmill. The problem is you. Right. And why? Because we're just, we don't like change. We don't like new systems, all that kind of stuff. And everybody who's sitting there listening to the sound of my voice and being like, that's not true. AI is great. It's just people have to be more curious and they have to just find their use cases. And I'm like, you've tried that, right? Has anybody responded to that in the same way? Has anybody responded when you've shouted, eat less and exercise to them? It doesn't work that way. When Microsoft is putting all of this effort, and I understand that there's nuances to this. It's not exactly like they took a pot of $2.5 billion and 6,000 people aren't putting into this. But what they're doing here, and I love Microsoft. I really do. I'm actually one of these people that I'm a huge Microsoft fan. But I keep telling them, you're the adult in the room here. You own enterprise. Why focus on it like this where you're like, yeah, let's come in and help you tweak the systems and work on the features. That's not the problem. The problem is that what they actually need the McKinsey's for and the BCG's and everybody else's is how to get people to sort of transform how they work. But even that doesn't work, Jaden. I promise I'm going to shut up after this. But even that doesn't work because the consulting model doesn't work. Why is that? Because when the big consulting firms, the McKinsey's, the BCG's, the Bain's, the Deloitte's, EY's, et cetera, ([Time 0:03:36](https://share.snipd.com/snip/05ee95db-0ccc-4db8-a96a-c60c274bf688)) - **Invest In Changing How People Work** - Do focus investment on changing employee behaviors and processes instead of only on product tweaks. - Connor recommends Microsoft and consultancies should prioritize coaching habit change over feature engineering. Transcript: Connor But even that doesn't work, Jaden. I promise I'm going to shut up after this. But even that doesn't work because the consulting model doesn't work. Why is that? Because when the big consulting firms, the McKinsey's, the BCG's, the Bain's, the Deloitte's, EY's, et cetera, when they go in, what they are great at is they are great at looking at A system of like, hey, you guys do sales this way. Well, we know that from doing sales in a million other companies, the best way is to do this way. So do it like that. Or you do operations this way, the best way is to actually do it like this. So change what you do. That's fine. That's great. That's called consulting. But coaching is very different. You can't just say to people, hey, be nicer to each other. Or hey, think differently. Or hey, be more collaborative. That doesn't come from best practices from other organizations. You have to get people to change their habits and behaviors and everything else. Which is why, sorry, now I've brought it all the way back, Jaden. Microsoft is trying to solve the easy problem and corporations are trying to solve the easy problem of like, oh, well, let's just tweak the system when the problem is not the treadmill. The problem is you just don't want to get on that treadmill. ([Time 0:05:17](https://share.snipd.com/snip/c0376667-39bf-496f-9e8c-73c4a1872f37)) - **The Individual Spark Drives Adoption** - Jaeden and Connor call the individual-level 'spark' the main driver of organic AI adoption inside companies. - The spark often appears through personal, high-value wins (time saved, creative boost) rather than prescribed work examples. Transcript: Jaeden Schafer The big change comes from individual people getting a spark and realizing, oh my gosh, look how capable this is. Look how much time I can buy back for myself personally to work on the parts of my job that I like doing. Automate the repetitive mundane parts. Like as soon as someone gets that spark, you don't need to spend $2.5 billion to convince them to use it. They're obsessed with it. I mean, most of you guys listening to this podcast are in that ballpark, right? Like you see, you see something incredible that it can do and you want to go all in and figure out all of the ways and even myself who covers ai all day every single day there is so many ways That i discover every day i'm like oh my gosh i never even thought of that that's so cool i want to try it out right like this is what drives the adoption is people being genuinely excited Because they they catch that spark and you have to ignite that spark for people and you're not going to do it by just going and telling everyone to go download copilot onto your computer And make sure to go ask it, you know, what time the Yankees are playing at XYZ, some simple little thing, right? Like, you really got to get them to actually get a spark with it. So that's one thing. And ([Time 0:06:53](https://share.snipd.com/snip/dabe49f3-ca04-4524-befc-e21882ba39c3)) - **Force Adoption With Process Not Just Use Cases** - Do create processes and expectations to force early AI use for those who don't naturally find the 'spark.' - Connor suggests embedding AI into daily workflows and making AI use an expectation, not optional encouragement. Transcript: Connor Yeah, I love that you call it the spark. So the thing that we figured out, I think, in this whole thing is the reason why we don't teach use cases and we teach process instead is that use cases, just encouragement. Come on, guys, find your use cases. Again, it's sort of like, come on, guys, don't you want to feel better? Like, get out there and run. Like, that doesn't work. You have to put processes in place. So for some people, you know, we do it through this whole behavioral thing, which actually drives a lot of spark because people like, oh, I didn't think of it like that. It's actually like this and not like that. So that gets a lot of people kind of like with that spark. But we can't get everybody. We just can't. So then the people that we don't get, we have to put a process in place. Right. So the two, let me attack those two sides. Right. So first of all, Gene, I totally agree on the spark. The only thing that drives AI adoption, it's one. First of all, it has to be driven at the individual level. Unless you're talking about AI in a product. Right. Like like Gene's AI box, right? Like that's your product keeps getting better and better and better because of AI, right? Like, I mean, and that's, that's awesome. But for individuals at a company, when you're trying to drive value, it has to happen at the individual level because you're paying people to drive value for your company. You don't have like some people in some robots, it's all people. So what do you do in that case? You want to have everybody have their spark. And the problem with the spark is that it often happens outside the scope of work. Anybody listening probably has had that spark, as you said, right? I want to know like how many of you have found that spark in a personal thing versus a work thing? Because I find it's often impersonal, right? You're like, oh, my gosh, I spilled coffee on this and I don't. And all of a sudden, Chad GBT gets you out of a bind. You're like, oh, my gosh. It doesn't usually come because somebody discovered how to write an email faster that who's that who's that like being like the genie has come out of the bottle right no it's like yeah It's not exciting it doesn't get you sort of like wanting to stay up all night like just working with ai it doesn't get those things so then how do you drive that if you if if you sort of like If people just aren't finding that spark right so? So again, like in our training, like we get a lot of people, but we don't get everybody. So the people we don't get, we have to have like the people leaders in those organizations, put processes in place, like meetings, things like that, where AI is integrated, and we have Specific ways of doing this, but like where AI, you can't get to the end of the day without going through AI in some way, because then it's sort of like forcing people. It's almost like, hey, listen, guys, to start the day at this company, we're all going to go out on a two mile walk or whatever. At some point, you're going to get more and more people who are like, you know what? I actually like, I'm really glad I'm doing this. You know what I mean? Like you force it in that kind of way of forcing mechanism, because then they will find their spark through that. But that's why we don't do use cases, because use cases are just like, come on, everybody just find your thing. And you're kind of encouraging X. So we say moving from encouragement to expectation of use, expectation of use is listen, all our meetings are on the third floor and the elevator doesn't work anymore. That's how you get people healthier. Do what I mean? Like you're to say like people have to go through a process. So that's how we've had success on that. ([Time 0:08:19](https://share.snipd.com/snip/7994694a-15bb-421b-a19c-f31e8b4c5543)) - **Combine Training With Mandated Workflows** - Try splitting efforts: ignite sparks via behavioral training and cover holdouts with enforced processes and leader-driven routines. - Connor explains training creates sparks for many, while mandated processes catch the rest and normalize AI use. Transcript: Connor You want to have everybody have their spark. And the problem with the spark is that it often happens outside the scope of work. Anybody listening probably has had that spark, as you said, right? I want to know like how many of you have found that spark in a personal thing versus a work thing? Because I find it's often impersonal, right? You're like, oh, my gosh, I spilled coffee on this and I don't. And all of a sudden, Chad GBT gets you out of a bind. You're like, oh, my gosh. It doesn't usually come because somebody discovered how to write an email faster that who's that who's that like being like the genie has come out of the bottle right no it's like yeah It's not exciting it doesn't get you sort of like wanting to stay up all night like just working with ai it doesn't get those things so then how do you drive that if you if if you sort of like If people just aren't finding that spark right so? So again, like in our training, like we get a lot of people, but we don't get everybody. So the people we don't get, we have to have like the people leaders in those organizations, put processes in place, like meetings, things like that, where AI is integrated, and we have Specific ways of doing this, but like where AI, you can't get to the end of the day without going through AI in some way, because then it's sort of like forcing people. It's almost like, hey, listen, guys, to start the day at this company, we're all going to go out on a two mile walk or whatever. At some point, you're going to get more and more people who are like, you know what? I actually like, I'm really glad I'm doing this. You know what I mean? Like you force it in that kind of way of forcing mechanism, because then they will find their spark through that. ([Time 0:09:24](https://share.snipd.com/snip/bc0cd3fc-ca0f-4da1-bcee-78828f92a31d)) - **Leadership Priority Determines AI Uptake** - Jaeden highlights that company priorities shape employee AI usage; if leadership doesn't prioritize AI, many skilled employees won't adopt it. - He contrasts fast-moving startups shipping biweekly features with entrenched companies doing 'business as usual.' Transcript: Jaeden Schafer Yeah. And I mean, the last thing that I'll say is there also is an issue in a lot of organizations where I feel like the companies themselves are not doing enough. And like you mentioned, that forcing function is so important because I feel like the companies themselves are not doing enough to encourage or push people. Maybe they had early like fans on AI. Maybe they just never encouraged it. Maybe they never rewarded it. And I mean, there's all sorts of absolutely ridiculous reward systems. Like who's using the most tokens. Okay. Terrible ideas, but, but like genuinely getting people to leverage these tools because the things I'm seeing a couple different things. Number one, a lot of the software companies I use and the scrappier, savvier startups, they are churning out new features every two weeks, because they just can and it's like, it's actually Blowing my mind. There's a bunch of companies I've followed for a long time. And the rate that they're able to just put out these incredible features or rebuild their entire platform is so impressive to me. And I understand why, because I'm doing the same thing with all my software and platforms and tools, but I'm not seeing it from all companies and the companies that I'm not seeing it from, For example, I mean, there's just so many people where you realize that we are sort of in a bubble. There's some people that are really taking advantage and there's some that are just not. I was recently talking to someone and he's a developer at like a big gaming studio. And he said, I'm like, I'm like, oh my gosh, like he's a front end developer. I'm like, you must just like Clyde must be your best friend, right? It's like so awesome because I'm doing this all day, right? And he's like, to be honest, like, I don't know if I've tried the Claude one yet. Like, I think I tried. We had like a GitHub one, but then it's we don't have it anymore. And like, I've started using it a little bit lately. But anyways, it just blew my mind. Much he didn't use it. And I'm like, if it's not a priority in your company, it's not a priority for your employees. And if it's not a priority for your employees, you're not shipping new updates every two weeks because you can, you're not 10 X in your output because you like, you can with AI. And so yes, there's like business as usual. ([Time 0:11:36](https://share.snipd.com/snip/ef2687ed-bc67-4219-9f58-8017191d4fc6)) - **Gaming Developer Who Barely Used AI** - Jaeden shares a developer story at a big gaming studio who barely used AI despite being a front-end engineer. - The example underscores the podcast's point that many professionals still aren't using AI even in relevant roles. Transcript: Jaeden Schafer There's some people that are really taking advantage and there's some that are just not. I was recently talking to someone and he's a developer at like a big gaming studio. And he said, I'm like, I'm like, oh my gosh, like he's a front end developer. I'm like, you must just like Clyde must be your best friend, right? It's like so awesome because I'm doing this all day, right? And he's like, to be honest, like, I don't know if I've tried the Claude one yet. Like, I think I tried. We had like a GitHub one, but then it's we don't have it anymore. And like, I've started using it a little bit lately. But anyways, it just blew my mind. Much he didn't use it. And I'm like, if it's not a priority in your company, it's not a priority for your employees. And if it's not a priority for your employees, you're not shipping new updates every two weeks because you can, you're not 10 X in your output because you like, you can with AI. And so yes, there's like business as usual. And some companies are like in these industries where you're kind of entrenched, you kind of have a good thing going. And business as usual works pretty good. But I mean, if you really want to stay competitive, you have to make this a priority. I know I'm speaking to the choir here, but it blows my mind because it probably happens at least once a week where I talk to someone and they're like not using AI. Connor This is the bubble we live in. It's the bubble we live in. Just to close that out, ([Time 0:12:36](https://share.snipd.com/snip/0935bdc7-1639-4e4c-8fa1-1f03cced9609))