Podcast

Man, if you run a small business, you are in all likelihood so overwhelmed by AI and the number of options that are out there. Most of us are rolling either Microsoft or Google for kind of the main environment for all of our tech. So why shouldn't you use the built in AI tools that they give you that they're pushing really hard right now? That's what we're talking about today. What's the right platform for most businesses? Who do I give access to? My AI champion, my AI team, everybody, by the end of today, you're going to have a clear framework for just what to do. Let's get into it. AI headlines abound like every single day. And it's so hard to make sense of this stuff because it all comes with zero context. Like, does this matter to me or does this not matter to me? We really need like a bigger volume of just high quality AI education and very contextual ways like AI education just for what you do. Because too often it's at like a 10,000 foot view and it's so hyperbolic, right? Or it's just through the lens of like, you know, a Microsoft MVP or something like that where they're just going to give you all the Microsoft stuff as much as influencers and all that. Like we're sick of being sold to online. As annoying as all that stuff is, there's a huge opportunity for the right version of education right now and I hope more folks get involved in it. Today I'm going to run you through kind of the seven leading platforms, why they may or may not be an option for you, the three like categories of folks that you would roll these out to within your business and a few thoughts on like how you, how you manage the life cycle of. Do people get access to those tools forever once you, once you get them set up on it. Because what you do after you give them access, just as important as what you give them. So let's cruise through the platforms. Number one is Microsoft's copilot ecosystem. The upside of this is if you're already on the Microsoft platform. Is it, Isn't that just nice? Isn't that where we want our AI assistant to work? Right. The worst version of AI is the one where there's a standalone assistant in every single app that I use and none of them can talk to any of the other apps. That's so irritating. The folks that continue to just plug in this standalone chat assistant into their end of their platform, like, nobody's calling for that, buddy. Like we don't all Want one more chat assistant to talk to. I think we get excited about them in very contextual use cases where it's like, oh yeah, actually just typing into this box is better than whatever the alternative was before. But nobody wakes up in the morning and thinks, I hope I get a, an AI assistant for my Fitbit today. Which is why Microsoft, I would argue, has crapped the bed harder than anybody else in the AI era because they should be like in the lead here. They made the massive investment in open AI early days they had. They had their hands on, on the lucky charms in a way that none of the other big tech leaders did. Yet Microsoft's AI implementation from inception, like has perpetually been so far behind of chat GPTs. And they'll tell you what is the same ultra powerful models behind the scenes. And that's just what makes it so much more shocking, so much more embarrassing that it's not better. Talk with anyone who has used both. In fact, it was over a year ago now. We did a YouTube video on the best AI for spreadsheets and we pulled in a few like dedicated spreadsheet platforms that had kind of AI built in. We tested a bunch of different AI assistants and wouldn't you know it, the best tool for spreadsheets that best AI assistant was ChatGPT. Literally having to give ChatGPT the entire spreadsheet or copy paste information in that was more powerful than the AI assistant that lives inside of Excel. So as much as I would love to give a raving endorsement of Microsoft Copilot in its current form, it still sucks. Like, it's still just not good. And the idea of it is fantastic and all the marketing materials are great and people can see that and read it on the surface and be like, oh yeah, that's a slam dunk. And it's a big time risk for business leaders who aren't actually like rolling up their sleeves and playing with this stuff to get sold on a bill of goods. Like the companies who committed hundreds of copilot licenses to their organizations. In fact, the very first rollout Microsoft did was to companies that had committed, I think, you know, 300 or more licenses or something like that. And they're paying, I don't know, 30 bucks a month or whatever the enterprise pricing is. Imagine that rollout inside of your organization where hundreds of people overnight get access to this co pilot and you're the big boss and you're like, okay, we just, we just spent a whole lot of money on this. Let's, let's figure out how to start getting that ROI back and you go out to actually use this with, with people and this was like a year ago when it first came out and it was even worse than it is now. And people actually start firing that thing up and they're like, this is the AI I've been hearing everything about because this doesn't kind of suck, it super sucks. Talk about getting egg on your face and wasting a whole pile of money. But the reality is if you actually use these tools, I've yet to talk to anybody who's like, oh yeah, the Microsoft AI stuff, it is so good. It just isn't. So even if you're on the Microsoft platform, still can't recommend it unfortunately. Now let's go to Google's equivalent Gemini. About a month ago they rolled out their Gemini 2 family of models and it's a massive Upgrade from the 1.5 generation of models and the 1 generation of models. A really, really big step. Google now has the strongest claim for like a, like a built in AI ecosystem kind of all in one sort of play. Now in practice I think where it falls apart right now is like that interoperability, like we want these magical experiences where I can come to a single assistant and be like, I, uh, what did I miss over the weekend? Summarize the most important emails and, and draft a few of them so that I can approve them. And right now like it just isn't integrated enough to be that useful unfortunately. And I don't know if that's just going to be a hard thing to do long term because that requires sifting through so much, I don't know, like such a volume of information. But for whatever reason like it, it's not quite the like, like Microsoft had great marketing around this when they first kind of proposed the idea of it. And this was there was a big marketing push in February of 2023. It was just a few months after the original ChatGPT came out and Microsoft's like gang, we did it. We gave OpenAI a whole crap ton of money and here's the deal, we're going to just have these AI co pilots that work across all of our apps for us. And you know, the Satya Nadellas and Bill Gates of the world were like, well yes, user interfaces, they're going away. We're just going to talk with our AI assistants and they're going to do our bidding for us. And this was the promise over two years ago now. And Microsoft, Google, still nobody has given us like in a business ecosystem a really good AI system that can work across all those different things. But, and this has changed in the last couple of months, Google is now a whole lot closer, believe it or not, than Microsoft. And the underlying Gemini 2 models are very, very good. Even if you're not a Google workspace nerd like you can just go to, I believe it's gemini.google.com and even with a personal Gmail account, you can test out a bunch of these models. If you're, if you're not on a paid Gemini plan, just recognize if you're on the free plan, your prompts are going to be trained into their models. It's not until you go to the paid, what do they call it, Gemini, Gemini plus or Gemini Premium. Whatever their paid version of Gemini is, at that point your data is no longer trained into the models. And so I can maybe kind of sort of see an argument for just sticking with the built in Google AI. Like that's the strongest argument I think you can make right now as an alternative to ChatGPT. But honestly, still third platform here, what I'm going to send 95% of small businesses to is the ChatGPT team plan. The team plan is the entry level business plan from ChatGPT where none of your prompts are trained into the models. You get some additional data security controls which for any small business I think you want. And there's a way you could like game this and try to put a bunch of the team on the free version. The problem is like you just don't have the same data security with the free version. And I as like a business leader don't want to have to rely on my team knowing what can and can't go into the AI model and just blindly trusting that. And in fact, in the early days of AI adoption, everybody was pushing for these AI usage policies. And honestly I think I'm kind of anti AI usage policy now because the underlying idea of the usage policy was it would define what apps you could use and what types of information you could put into those apps. But for me, I don't know that I even want to put an app in front of my team if there's like clearly inbounds and out of bounds ways to use it. Like I just want an app that is locked down enough to where they can't accidentally screw that up. And on the free plan of ChatGPT, like we shouldn't be putting identifying client information or a lot of sensitive stuff into those apps. And it still is funny to me that like we bulk at the $30 a month price tag for ChatGPT team, like, the, the amount of value I get out of it is just like wild and like I can go and I can get a room full of people very excited about the future of AI and how it's already changing things. And it's, it's not just hypothetical. It's here are the extremely exciting things that we can do today. And they're like, yeah, no, I could totally see this change in the future at work. And then I'm like, it's 30 bucks a month. They're like, I don't know. That kind of feels like across nine people, that's how much money I'm like, people, this is a big deal. In fact, I heard it framed recently. As someone said, basically I'm no longer willing to employ people who aren't capable of making good use of AI assistance. Like it has risen to that level. And I can tell you the delta in my productivity with and without an AI assistant is very big now. I don't like, we are far from the majority of the workforce like having that level of skill to where it's making them much more productive at this point. There's for sure outliers that know that stuff well and are really putting it to work. But what isn't going to solve that is withholding this stuff from our teams. The reality is if you've been paying somebody to work for you up until today and you haven't put AI tools in their hands, that person is now a bit hamstrung by not having developed this skill set yet. But what are you going to do? I'm not going to cut that person and then go hire somebody that already has amazing AI skills. Like I have to upskill my workforce. And the only platform that has been at the top or near the top since the inception of AI has been, I will say, since the inception of like the modern generative AI age. As kicked off by ChatGPT, the only platform that has been at or near the top throughout has been ChatGPT. And it comes with a bunch of really nice to have features like stuff people don't talk much about. Canvas, Very helpful for writing data Analyst. Its ability to write and execute Python code to process like large amounts of data, custom GPTs that I can share with my team just comes out of the box with a lot of really handy things and it's not this big complex thing to learn. Like you look at the Microsoft ecosystem, Land of lakes, the number of AI SKUs there are like you don't even know where to tell somebody to go. You to begin. Whereas ChatGPT is just like you sit down and there's two types of people in the world. There's a type of person that puts their name into it immediately and sees what it knows about them. And then there's the person that puts the hardest thing that they know, that the really technical thing they just figured out, they put that into the AI first because it's the. It's the best marker they have of is this as intelligent as me or not? And when it doesn't get it right, they're like, neener, neener. Still got it. There's. You're one of those two people. The first time you sat down in front of AI, same as the first time you were sat down in front of Google. Two types of humans on this earth. That's it. Today's episode is sponsored in part by TeamUp, the folks that help you hire people directly in the Philippines to work for your accounting firm. TeamUp's put together a killer guide, actually a 2025 salary guide for hiring accountants in the Philippines that you can swipe for free down in the show notes. One thing they called out in this guy that I wanted to cover today, what it means to direct hire. Because a lot of offshore groups, you're actually hiring the company and they have an office and they have kind of this resource pool of people that are getting work done. Team Up. Their approach, you hire those people directly. So Team up pays recruiters on the ground in the Philippines to find those people to come and work for you, which means they're just another member of the team. If you run a distributed team, they are no different than the folks that you got across town. Now, there's sometimes like, disclosure rules, stuff like that that you got to navigate, but in terms of, like, building a workplace environment, they're just part of the squad. It means there's no, like, middleman that can complicate that relationship or anything like that. It means the culture and the relationship that you have with them, it is not entangled in any other stuff that's going on, you know, in their office or, you know, the aspects of their work experience that you don't have control over. Like, it's just you hiring a person to do work. Simple as that. The salary guide. Super good. Grab that down in the show notes. And thanks to Team up for sponsoring this episode, gang. Today's episode is sponsored in part by the fine folks at Carbon. Every year, Carbon does a report about AI and its impact on the accounting industry. They survey a whole bunch of accounting firms. You may have heard. We actually did a whole episode of the podcast about it. It's a really, really good report. Bunch of interesting insights from real accounting firms to see how they're using it, what's working, what's not working, and kind of get the commentary on either ends of the spectrum. A lot of really interesting trends that developed just in 2025. A whole lot less fear in 25 than there was in 24. People seeing it more as an opportunity. Still a very big divide in the excitement of leadership versus staff level folks. That means we as leaders need to be doing a better job communicating and normalizing AI as a part of work so that it's not this amorphous, spooky thing that folks kind of are unsure of. Lots of exciting stuff happening there. I'll link access to that AI report down in the show Notes. Totally worth checking out. I think you like it. So chatgpt team. It's in my opinion like it's the right answer for 95% of small businesses. I'll cruise through the rest of the big platforms. You hear about number four, Claude. Very, very good. It has probably the best writing model 3.7 sonnet just recently got basically Internet access. Kind of amazing that up until now you had no way of using these Claude models with like up to date information from the Internet. But it doesn't come with a lot of the features that are really useful to me that I get with Chat GPT, like Deep Research is a huge one now that that's gotten really, really useful. And so while Claude's got some kind of cool party tricks, if I'm going to put something in the hands of my team, I'm still going to go chatgpt first. 5th Platform Grok. Grok 3 is an excellent model. Very, very good. Performs well on the benchmarks, but as a complete app is Nowhere near what ChatGPT is if I'm trying to use it for productivity with my team. Number six Perplexity. Unbelievably good AI search platform. And they're doing some cool stuff now, especially on their enterprise plan, which is $40 a month where they can connect to your Microsoft or Google workspace and basically let you do perplexity search off of your own company data. Which is really freaking cool. Because if there's one thing Perplexity probably does better than anybody else right now, it's giving you high quality answers based on information that it can go out and find and if you think about all the, you think about all the valuable company stuff, be it client files, you know, a file folder with everything you need for a specific project, a really good AI that can see into all of that stuff, that's pretty juicy. But at the end of the day, perplexity. It's still really just a search platform. It stops short of like doing much more than finding for me. So it's got its place. But I wouldn't put it ahead of if you're going to like, if you're going to get out the checkbook, I wouldn't put it ahead of ChatGPT. And the last platform number seven is the long tail of something else. And there's a kajillion AI apps out there and behind the scenes they're using the same AI models as the platforms we just talked about, by and large. But if somebody runs into your office one day and says, oh, I found this, this amazing AI assistant that's kind of off the beaten path and we should go chuck a bunch of stuff in it and use it for this, like, pump the brakes. Because these things are like spinning up and going away really, really fast right now. These standalone AI assistants for very specific jobs. And you got to have a really good understanding of like the data security behind them to ensure that that's something that you actually want to pull into your business. I don't know, I'm like an early adopter, but in some ways also a late adopter. It takes a lot for me to go off like the, the, the menu of what's kind of the norm. And like these platforms that we've talked about, Microsoft Copilot, the Google, Gemini stuff, ChatGPT, Claude, like these are very mature platforms where we know what we're getting from them and they're not like a startup shipping a brand new thing, launching a new thing. I cannot imagine a version of that brand new thing that would change my mind to say, no, let's pull the plug on ChatGPT or Gemini or something else. Unless it's doing like a very, very specific job in a silo. So in some cases that will be okay, like maybe it's going to help me draft my LinkedIn posts or something like that. Like it's solving a, a very specific pain that we have. Just be mindful of the information that you're putting into it, security considerations and all that. But that doesn't ultimately replace what are these really good sort of universal AI platforms that we have now, three types of users. Because even if you buy into a platform, you got to decide whose hands am I going to put it into. First type of user is your AI champion. And this is, this is your big nerd. Everybody. Like, it's like a really big spectrum of how people take to AI assistance. There's some folks who sit down and they're overwhelmed by it and they're like, can somebody tell me what to do? There's other people who sit down in front of that empty box and they're like, where has this been my whole life? I know exactly what to do right now. And then the first person looks at that and they're like, I don't, I can't even understand how they would use that. Your AI champion, the person who is probably most vocal and most experimental in the ways that they're using it, they need to be the person that is wired to be really excited about this stuff. Like, folks are at their best when they're geeked. Being able to geek on something, it is a, it's a gift because they're going to think about it when they get home at night, like, whether it's their job or not. They're just into this stuff. And it's those people who will always be best at that thing. That AI champion, there's an argument you can make to actually put several tools in their hands, because if there is an off the shelf tool or, you know, Gemini or Microsoft Copilot or something a little more different that is very valuable to your business, that is the person in your firm that ought to be discovering those sort of fringe use cases. Like, there's enough happening in AI to make space for experimentation. All the normal folks on your team who are not maybe enthusiasts at this, they're not going to like Excel at this sort of experimental work. But your champion, there's an argument that you could put several of these tools in their hands, probably on like limited time horizons, where it's like, okay, we're going to try this tool for 30 days because we have a specific idea of what we could do with this. Take Perplexity, for example. You can now connect that to, you know, specific folders in your workspace. We want to test if this will let us do a better version of search across our client files, for example. And so your AI champion will spin up that subscription for a month. They'll do a bunch of testing on it. You document that testing and then you decide to either do a bigger rollout or stop. And it's very important that you document what you learned and why it didn't work along the way because in all likelihood you're going to return to the same possible use case down the road, either with the same tool or with a different one. That I think is the best way to use our AI champion is when somebody gets a good idea that's a little bit off, off the normal menu, we throw that at a dedicated person that has wider access to tools. Second type of user, my AI team, the squad that has been called in to kind of, I don't know, lead the charge. I have mixed emotions about dedicated AI teams because you run the risk with a team of making it look like it's not everybody's job. If we say, and we ran into this with RPA robotic process automation, it was really hot for a while and you would have kind of dedicated groups that were the RPA experts. But then because they didn't have the domain expertise and didn't understand the underlying problems, they would have to like go and sit with an actual user in order to understand what was going on. And then the rest of the team, like, honestly, the rest of the team puts themselves at risk by not learning anything about RA rpa. That is because it's somebody else's job. And so as soon as you make this thing like someone else's job, there's this implied notion that it's for someone else to be worrying about and not about you. And AI is to the level of importance to where I think it needs to be, sitting on everybody's desk, something that everybody is using. So third type of user is just anyone and everyone within your business. And for me, today, I'm going to put ChatGPT team into the hands of everyone. And for the AI team or the AI champion above that, we will experiment with like a broader set of tools. But the bare minimum that I think everyone needs to have their hands on today is ChatGPT team. I really believe that, like 2025 was a big turning point with the new models that we got at the end of last year. And then deep research is another hugely beneficial tool for 2025 that I honestly like. This sounds very hyperbolic. I try to avoid talking like this when it comes to AI, but I don't know that there's any type of knowledge work that should not be testing deep research to make them more productive. Like, it is incredibly good right now. On the team plan, you get 10 deep research queries per month. I hope that goes up over time. On the $200 a month pro plan, you get over 100, but you don't get the same data security with the pro plan. So, like, if I'm going to use it with everyone in my firm, not a chance. I'm putting everybody on the pro plan just due to cost and security. But right now, in the team plan, each person in your team is limited to 10 queries per month, which is kind of a bummer. But I really think in 25, we hit this tipping point where AI is like, good enough to like, save the majority of people a pretty significant amount of time. In the lead up, like 2024 and earlier, there were a lot of people that it could save time, but not everybody. And it was a little bit more of an exercise of we can see where this is headed. We're pretty sure soon it will be significant. And so now we're readying ourselves because the people today that are crushing it with AI, they were the ones that had, like, already rolled their sleeves up a year or two ago. We're already comfortable with it. And so when this new stuff hit in 2025, they're like, holy smokes. This is just like pouring gasoline on the fire for me. And whether that was you or not, whether you've like started with this at all or not, the best time to start is this old thing. The best time to start was a year ago. The second best time is today. And it's a skill set that is like becoming essential to work the skill because it is a learned skill of being able to get the most out of a chat assistant for the same, like, many of the same reasons that people will get varying levels of utility out of colleagues, out of assistance. Like, there's a skill set to doing it. Delegation, like being able to get out of the way, documenting what you do in a very similar way working with AI assistants, there's things that are going to help you and things that aren't going to help you. And you got to learn that nobody's going to like, get out of bed one day and just know how to push it to the limit. Today's episode is sponsored in part by cloud account staffing. They will help you find super talented accountants in the Philippines. But they're not just going to drop off after off the map after that. They will help you with a lot of the logistical issues that come with hiring folks in a whole other country. How do I get them a computer? Like, very basic it things like that. What happens if. If they wake up one day and their computer doesn't work? How do they get a replacement? How about power outages? I Was talking with the cloud account staffing folks the other day that like they actually will help the folks that they pull into firms like have like backup generators, like whatever they need to have a redundant power supply. Redundant. Redundant. That's not the right word, you know what I mean? Cloud account staffing, what they do goes well beyond just finding smart people to come and work for your firm because the reality is there's kind of other logistics that come with that, some easier to solve than others. And I need help, I do. And cloud account staffing wants to make that as easy as possible for you to learn more about that, what they can do for you. Check out the link down in the show notes. This episode is sponsored in part by Makers Hub gang. It has never been more important than it is today to ensure that you have hitched your accounting firm wagon to the tech partners that are going to take your hand and lead you into the AI era. I don't care if you think AI is gonna gonna be the end of everything or the start of something better. AI is already here, it's already happening. And Maker's Hub is at like the bleeding edge of what AI can enable for the accounts payable process. I know, wow, sexy. Listen, just be happy that they're doing it so you don't have to. They're using AI vision models and LLMs to push, you know, automated extraction from bills and kind of the subsequent logic further than we've ever been able to do before with just pure OCR technology. And honestly the bigger thing is when you have somebody that knows their stuff like that. And Makers Hub has come around really just in the last couple of years. So really in the age of generative AI, I have a high degree of confidence that whatever is around the corner when it comes to AI, they are in the driver's seat to lean into that stuff in a way that oftentimes the incumbents are not Makers up. You're also gonna see them out at a whole bunch of accounting conferences this year. So they are having conversations with accountants. They understand the problem of the firm, how you run a scalable bill pay service. So if that's something that your firm does or if you got clients with gnarly bill pay issues, makershub do it this time. Platform of choice. Learn more about it down in the show notes. Now, last thing to think about is what do you do after you give your team members usage to this? I've seen two different approaches. One, because a lot of people have fear they're going to give this to their teams. And it just, it won't change behavior at all. Right. One approach, use it or lose it. So every 30 days, if that person's using the tool, they can keep it. If that person's no longer using the tool, you take it away and it's like, well, you kind of had your chance with it. If you're not getting any mileage out of it, no harm, no file, like we're just going to cancel that subscription. Second approach is mandatory reporting, like a mandatory feedback loop of how you're using it. And this is to like increase organizational learning while driving some accountability. So if we go back to the person that has the mindset where they're like in 2025, I'm not even willing to employ somebody that is not capable of making use of AI assistance to be more productive. If we go to that person's mindset, then if we think about the use it or lose it plan, can we really make it optional for people? What if they don't use it? You cancel that subscription. You've now got this person on your payroll. As AI developments continue to come thick and fast, these tools get better and better. We find new and exciting ways to use them. You now have this one person on your payroll who didn't use it, so they're not learning along with everybody else. I could see that maybe making sense last year in 2024. To me it feels like I'm increasingly feeling more like that person that's like I'm kind of just not willing to employ somebody who isn't making good use of these things. I'm leaning more into that camp, which makes the whole mandatory reporting of how you're using it, it makes that feel a little more essential. And that may seem extreme. Like am I missing something like being overly dramatic here about how big of an impact this is having? I actually, I was in, in New York last week giving a talk, chatted with someone from a pretty good sized accounting firm that said like this year everybody has to report back every single week how they're making use of AI. Like it's risen to the level where it's, it's not optional anymore because they see such opportunity there. And my hope is that we're like past the hyp of AI. Like I would love to. You know, it's not new now. It's, we're almost two and a half years post ChatGPT. There's all sorts of like new different hype cycles like vibe coding and a lot, a lot of that stuff right now. But My hope is that if we were going to force AI on a business too soon on our team members, that it would happen by now. And if we're really settling into this new reality of our expectation of the team is that you're making use of this stuff and we want to hear about how you're doing it. I'm hearing this more and more in the wild. And my hope is that, like, this is a more thought out, considered, like, understanding of AI. Like, we're not pushing this because we're seeing the news cycle and thinking it's a big deal. We're pushing it because we've had team members find such success that that success has kind of risen to the level of being essential for the rest of the team to lean into. That's the impression I'm getting. I mean that at my core, that's how I feel 100% in the ways that I use AI every day. But talking to more people who will leave it up on a monitor. So, like, if you are an accountant, you got the mega battle station, however many monitors you have, the worst thing that we can do is leave something up on there that can potentially distract us and take us out of deep work. Like, but most of us, like, we've got Outlook up, we've got Slack or Twitter or something else up, just like inviting distraction, right? And the biggest thing that keeps us from using ChatGPT is usually just remembering it exists. Like having to open up that tab and go out there. And it's why, honestly, I recommend the desktop app because then it can. ChatGPT can always be just a keyboard shortcut away. But if you got that extra monitor, especially if it's showing something that just perpetually distracts you, man, put ChatGPT up there. It will just remind you that it can be there to help you for all sorts of things from, like, ideation to just getting unstuck on something, to processing a big, like, file of data from one format to another, to writing an Excel macro, to extracting useful things from a meeting transcript, or pull in 10 meeting transcripts and have it kind of build a framework of how you advise your clients. That's actually a really interesting one. Go to Deep Research. And I did this with myself the other day. I've done it with a few other folks who create content online, where I say, go out, find all the information you can from this specific individual and then based on what you see in their content, create a framework for how they advise folks on xyz, I've done this with Myself and it comes back with really, really high quality content because it will go to the podcast and the YouTube, pull transcripts, the stuff I post on LinkedIn, Twitter, my bio, like it rolls all this stuff together. And Deep Research, which works for 10 plus minutes at a time and will pull upwards of 100 plus sources, it reasons through all of this stuff, like way more work than I could do in 10 minutes time. And the result you get back is actually really freaking good. Like the notion like there's a world where I could. And we have this, the AI assistant that is trained on all the episodes of the podcast. Honestly, there's part of me that's like, I don't know if this is necessary anymore because chatgpt with web searching enabled and specifically Deep Research has gotten so good that if you have a question for me, if you chuck that into Deep Research and say, pull on like the entirety of Jason's stats, like the accounting, thought leaders content to give me an answer to this. I honestly don't know that I could do anything better or more complete than that answer. Now we'll see how it does. Like when you give it specific context, like I'm going from this tool to that tool or I have this very specific situation. That's where I think I've probably got it beat. But that only works if you've got me live like one on one. Right. I've just, I'm kind of blown away at how good it is. So three things. Decide on a platform, what's the one platform you're going to give to everybody? And then consider if there's any other experimental platforms that you ought to put in the hands of your AI champion. Two, decide who your types of users are, from your AI champion to your AI team to your everybody and who gets access to what. And then three, is this like a use it or lose it type program? Or do I have an expectation of how folks will use this? And have I created a space for folks to share wins and like enable that feedback loop? Sometimes a problem with like the biggest problem I see with tech rollout, where leadership will come out and be like, look at this cool thing that we're super excited about and you should love it too. And then you put it in front of your team and they're like, yeah, I don't, I'm not that excited about it. And you're like, what do you mean? How are you not excited? Can't you see how it saves time on X, Y, Z and you're talking to a salaried person who has no incentive than being more productive. I would love for it to be in their best interest to want to make the whole company produce more. And I was wired that way. Maybe you were wired that way. But if somebody's just making money with, like, their butt in a chair, they're not really incentivized to be more productive. So also consider, what are the incentives that they have to be higher output. If those incentives are better aligned with your incentives, then you're probably going to see a little more excitement and curiosity around pulling in tooling that will make them more productive. It's a lot to bite off, honestly. And the biggest mistake we make is overthinking this and overanalyzing this and thinking that we need this big grandiose plan. You kind of just got to start, because until you start, you won't. You won't begin learning. Any tips you've got. Like, if you have made this bigger rollout with your team, stuff you got right, stuff you got wrong, drop that stuff down in the comments. It'll save people some headaches. You disagree with any of this, you got an alternative platform that you prefer, let me know and I'll see you in the next one.