022 Big Four Thoughts And More!

You got Q's, I got A's. It is Q and A. Roundup Wednesday. Doing a little check in State of the Union, State of the Daily Show. But we're gonna be covering some spicy topics from how do I feel about the big Four? To how do I heck do I choose a tech stack when all this AI stuff is happening? Hector Garcia just made a robot of himself that we can all go out and have a play with. We're gonna run through it all today on this special Wednesday Roundup edition of Jason D. Okay, can we just. Can we all just take a moment, check in on the state of the show here? Why do you all keep coming back every day? I'm just. I'm like, this show is far exceeding my expectations. You know what I think it is? I think it is a great example of the value of, like, doubling down on the people you're most connected with. So if you got a hundred clients out there, but you got 10 that you just, you love, you jive well with, is there something else you can do to, like, enable those 10 clients or to go deeper with them? And that was part of what excited me about the whole idea of this show is, like, the folks I best connect with, what would it mean to just do more with them and connecting more of those people, too? Like, we've got all sorts of stuff going on in the comments every day now, and that's just been a lot of fun. This thing, like, it's first 30 days. We're at almost as much watch time in the last 30 days as there was on my main YouTube channel, which was, like, working on that for like, two or three years. Invested outrageous amounts of time and money into that. So thanks, y', all for being here. If you're still getting value out of this, that's awesome. Tell a friend. I don't know, y' all have been, like, talking about this online and stuff, and I think that's how most people are finding it. So super appreciate that. I never. When I. So I never, like, totally expected this thing to actually be a daily show, which I realize now is unclear given the name. I just thought it'd be more frequent than, like, the main channel videos where I posted like, once a week. This may not forever be like an every single day sort of thing. It's just more, like, unstructured, more frequent than the other stuff. But here we are, and I'm having a blast. And that's why I'm still doing it every single day. And people are like, how are you gonna have something to talk about every single day. I'm not gonna lie to you. I got like, 10 solid discussion things I want to cover for, like, future episodes. But then in the meantime, there's always new stuff that comes up, like Kanopy's release yesterday, stuff like that. I just think there's a lot to talk about right now, and a lot of it comes from y'. All. Like we're doing today, sharing ideas and, like, thinking through new stuff. So let's run through some Q&As first. Okay. DM, somebody sent me on LinkedIn. I don't know that I'm allowed to share who they are. They said I started putting out the smallest bit of content on Facebook for my new firm. And, oh, my God, I had to Google what the word serendipity meant when you first said it. And now I'm starting to live it. Hopefully I'll be able to go full time soon. You kidding me? You stop that. How exciting is that? It's true. You gotta start putting yourself out there online. You know, there's like a shocking percentage of people who are here who are just lurking. Stop it. Leave your first comment below. You weirdo. Stop lurking and be part of the club. We're all gonna get a little better together. Okay, Question from Herman. Kind of a question. He said, another thought. How will AI enhance integration platforms like Zapier or Make if you have a trigger that can reason and decide actions based on learned context and not just like, very narrow scenarios. So this is referring back to the episode we did on Monday 20, which can I say was an absolute banger. If you haven't heard that one yet, go back and listen about how kind of context is like the currency of the future. Gosh, you ever say something and then you just think, gosh, you sound like an such a nerd saying that. But it's a good point. So he's talking kind of about AI agents who can sort of go out and perform their own functions. And ultimately, the most useful AI agents will be those that can perform functions based on instructions that you've already provided. So if you have a clear SOP that you've developed for doing a certain thing, when we get AI agents that we can use ourselves, which we kind of already have in very basic versions, the firms who have well documented procedures, they're gonna win out. Because just like human users can follow those procedures, AI agents can also follow those arguably in a much more granular way, because they can also go out and, like, look up all the help documentation to do those things and all the stuff that our human lazy butts are just too lazy to do. But where does that leave? Like Zapier, for example. And I'll be honest, I replied to you initially on this question, Herman, and like, I've waffled back and forth several times on that. What I don't know is if AI's ability to do this replaces Zapier. Because the future of custom automations is just semantically describing what you want to happen. Like, that's just where we're headed. We'll probably have that by the end of the year. So, for example, let's say I've got a new client form on my website and that's an airtable form, and I want to create a zap to email myself every time somebody completes that new client form or post a message to a Slack group channel or something like that. And so that's my zap. It's a trigger that a form's been completed and then an action to send me an email or send the Slack message. I can go to AI and say, in fact, you can even try this in ChatGPT right now. And it doesn't create an API call to send me an email every single time that this form gets completed in airtable. It can see all the API documentation out there and work out how to do all of that. Right now, ChatGPT exists like within a box. So there's no like, persistence or way to create that automation to run on a recurring basis. But AI agents and so like Auto GPT, baby AGI, they can run and execute like, they can write and execute their own code. So all that is to say, it's not going to be long before an AI agent, you can just semantically say, I want you to do this thing and it creates that automation. But does that replace zapier or is that just the next version of zapier? So when Chat GPT announced their plugins a few weeks ago, if you haven't heard, there's. It's like a way to extend ChatGPT like a whole ChatGPT ecosystem, similar to how QuickBooks has its own ecosystem. Developers can now develop for ChatGPT to like, read and write data from within ChatGPT to your QuickBooks to any of the other apps that you use. And one of the first ones they mentioned in that announcement, There were like eight or 10 plugins that people got in early access was zapier. And it was like, oh mama. Like, does that mean I can now chatgpt with all the stuff that my Zapier is connected with, that seems pretty cool. But ultimately I don't, I can't decide whether AI is a replacement for that thing or whether that's just the next version of zapier. So if I can semantically describe the zap that I want create that automation in Zapier, absolutely. That's just a way better version of that product. Right? Because we all get stuck like fussing through all these fiddly little aspects of how it works. But ultimately, if you're setting automations up to automatically run, they have to live somewhere, you have to go be able to manage them and change the frequency and check to see if it ran last Tuesday and stuff like that. So getting back to the question, does it change how these platforms work? One way or another, the platforms are gonna change in a massive way to enable the ability to create automations like semantically and Zapier and make have a, a big advantage there because they already have the underlying infrastructure to support it. But whether that replaces them or makes those tools better ultimately, I don't know. It's probably too early to call. I will say on the other side of all this, we're just in a fundamentally better place because you can explain what you want that automation to be rather than having to go out and learn a new tool and connect your airtable and connect your email account and all these different things that are just kind of generally a pain. So all that is to say, heck if I know, but I just know that whatever, that next version is going to be better. Okay. Mo said. I like the idea of teaching it, it being AI what is on my mind and would be cool if it can help me in the future when I make decisions and it can remind me what my thoughts were before and possibly protect me from making mistakes or errors that were made in the past. That feels relatable. So I'm real hot on the idea of like a kind of a personal journaling AI. A lot of us are verbal processors and the idea of doing like a brain dump for 5 or 10 minutes at the end of the day of everything that happened and then AI like structuring that information and it knows, okay, all your work stuff happens in ClickUp. Based on everything you just told me, here's some to do's that probably ought to get parked into ClickUp. This seems like a personal thing. You manage your notion, your personal stuff and notion. I'm going to chuck a few of these things into maybe your notion capture system. So you've Got them somewhere. But also then just generally that AI getting more helpful over time as it learns your various projects, the things that you work on, your thoughts on different things. Because if it has all that history and I'm going to chuck a new thing in there, that response is going to be fundamentally more helpful than something that doesn't have all that history. Right. So something with long term memory that's going to get more helpful over time. I'm excited for that. I'm actually working on a version of that. You can absolutely build it today. You kind of got to hodgepodge several things together. There will be people like companies that nail this, this kind of like, kind of like a AI diary sort of thing and where that assistant on the other side then gets really helpful. Honestly, like I like if you've seen the movie Her, I see kind of some weird stuff happening over time with how that learns you and like you pouring more and more information into that and like what those interactions look like. But another issue that sort of arises is you now have all this context from your journaling that is like lives in a single place and it is its own chat and voice experience. But can it see into everything else that you do? I imagine in like if you'd shown somebody in 1995 the way that we use the Internet today, they're like, oh cool, I can see that this happens within an Internet browser. And you're like, oh no. Actually what I do is I use like 60 apps every single day. They're going to be like, you kidding me? I use WordPerfect and Play Asteroid and like those are the two apps that I have. I see us heading towards like a, an issue of having all these chat and voice interfaces that are segregated, siloed. You don't have all that context. Like it's not connected. And maybe that's part of the play that Chat GPT is trying to make with plugins is that then becomes the interface for a whole bunch of things that we use, but we just use them through ChatGPT. So I love the idea of like the personal journaling stuff. Where that gets messy though is like if I'm gonna do a brain dump, it's gonna be about work stuff, it's gonna be about personal stuff, it's gonna be about that birthday gift I gotta get Steve and it's gonna be a reminder that I need next week. Where should all those things go? I don't know. Still some stuff to figure out there, but still seems like pretty darn cool if you think about like version one of like Google Assistant and Siri and all these things, man, three years from now, like, and don't get me started on Siri, which is just so inferior to Google Assistant. It's just so irritating. These assistants, even the best ones today, like, they're gonna look so antiquated in a few years time. Okay, Smudpie, you've never mentioned Taxdome at all. One of my partners, the president, is hot on it. I'm just not sure. Do you know where they are in this whole AI vision? Not really. I have talked with a good number of people and I just like, I can't like call out anybody specifically, but I will say everybody's in different places on their AI journeys. I have talked with folks who are like disappointingly behind, who are now doing great. And so like, even that's kind of a moving target. Ultimately, I think we got to reward the people who are actually shipping stuff because it's one thing to sit back and wait and see, but the people who are learning the most are the people who are shipping stuff right now. You're not going to like build something for two years and then ship it and be like, ta da, we got it perfect. Like, that's just not how life works. I will say, like, I'm impressed by the people who are now iteratively, Iteratively shipping things because there is some risk and being the first to move on that stuff. But that's ultimately how you learn. And so two years down the road, the people who will know the most will be the people who have actually been doing it. I don't know if Taxdome shipped any AI stuff. Nothing that I've seen, but maybe I missed it. But people are also like, I mean, he used to have like a roadmap for, you know, the next 18 months, and that was kind of your plan. People have also like just completely thrown that stuff out the window with how fast everything is changing right now. Okay, Hector Garcia. There's another callback to the episode from Monday where we talk about building bots on your own context. He built Hector Bot. I'll put a link in the. In the video description and the podcast. What is the show Notes? Gosh, you can go take it for a spin yourself. Hector bot expert in QuickBooks Desktop so Hector's an example of somebody that's got a wealth of context, right? He's been shipping videos since Lord knows when, training on QuickBooks on all sorts of different things. But that context is now an asset. So is that chatbot or that access to that context. Is that something that you give away for free or is that the most pure essence of Hector? Is that like a Hector in your pocket? How much would he have to pay to do a one on one meeting with Hector right now? I don't know, a lot of money. So if you have like chatbot access to all of his context, does that become a really expensive thing or like a. I think I use the analogy of like the sawdust, like the byproduct of what you've already done. So you may as well just give it away. We're talking about Hector here, but that applies to like all of us. All of us with personal expertise. This is a thing that is absolutely monetizable. It feels a little sci fi right now. And like you kind of feel odd, like putting out this thing that says, hey, you pay for my bot and it's X dollars per month. But in the future, when ChatGPT can see across everything on the web, that's what you got left is like your personal context, what you believe is correct. I was talking with Matt Metrus about this the other day, by the way. Matt Metris a big old crypto expert, and I said, I want like a mat. Like I want my own mat. And Matt goes around doing talks on crypto, taxation and all these things, and he knows all of those web three words that most of us don't understand, like staking and all these other confusing things. And you can absolutely wade through all the rules that are out there and the rules that aren't out there, where there's like ambiguity on certain things. Or Matt can take the context from all of these talks that he's given, all the source material that he thinks is relevant, and put that all in a single place. So I've got Matt in a box and his expertise, he can continue to enrich that body of cont. And I've got like my own man on speed dial. It gives citations to like, where that stuff is coming from. I mean, like, talk about the ultimate high leverage use of domain expertise. Right. I love that. That's just that, like, that's super exciting to me. Who, Akshay? Yeah. Tell me the right way to say that. It says, curious to see the tech platforms that can best leverage. Ooh, Advisor in the loop. See that as a massive opportunity that can combine the best of AI with what we do as accountants. Absolutely. So you got something like ChatGPT right now, but let's say you strap that to a QuickBooks file and your client can just go ham talking with the QuickBooks file all day long. If you put a tool like that in front of accountants and attach that to say a month end financial statement delivery or a tax return, the pro is probably not just going to let the client go wild and ask a whole bunch of questions without being able to like validate that those answers are correct. But man, clients would absolutely love that. Imagine being able to just have a chatgpt that could see a tax return and they could sit there and ask questions of it all day long. Why is box this? Why is box 14 this? Like I, this number isn't anywhere in my information. Oh, it's actually a combination of xyz, you can look on page this and that. I think clients would love that because they wouldn't be afraid to ask dumb questions. They wouldn't be afraid of asking the accountant a bunch of questions about and bothering them and possibly racking up a bigger bill. It would be awesome. But for an advisor that's like there's an aspect of liability there that maybe we aren't excited to get on board with when the AI hallucinates. So you're probably not just gonna send them like cut them totally loose to go do all that stuff willy nilly on their own. But where I think there is a cool opportunity and like to be clear, I wouldn't be surprised if that thing already existed. A tool where you can upload your tax return and then you can, you can talk with it like a consumer facing tool or upload a set of financials or connect your QuickBooks and then talk with your QuickBooks. In fact, genius Sheets has already got the Talk with your QuickBooks thing. But ultimately how can you marry like the expertise of an advisor and like the accountability of the advisor with a chat experience like that? It's some like, it's some hybrid of asking the accountant and like this kind of solo chat experience. So for me that looks like something that is like the accountant has control over like the level of confidence the AI has to have to just give a response. So let's say the AI is 90% sure. Great, you can go ahead and just reply to the client immediately. I can still see a log of those conversations, but if it's anything below that, maybe I need to get pulled in. But then when you get, you know, notified that hey, can you come help answer this question of the client? The AI can then fetch all the information it thinks I need to be able to answer the question, maybe even suggest an answer and say, does this look right? According to all of this information that's already helpful for me, that's already better than a cold email that I get from a client without any of that context where I have to go out and pull all that information up myself. I love the idea of that. Like, that seems like the ultimate marriage of expertise. And what I will enable that will be really interesting to see. Like, obviously that's like, not to cruise past how complex that is and how maybe you don't know how you should feel about that and who ultimately becomes liable if, like, the AI does lie about something. It's a. It's a complex calculation. But I know that, like, our clients, like the people on the other side of it, I think, would absolutely love something like that. Okay, cj, I know you like ruffling the feathers from time to time. Pasha, can I make a request for a problem with the big four slash live firm environment video? Okay, well, your questions lit literally not a question. It's. It's already saying how you feel about it. So I'm not sure how that much of that is a question for me. A problem with big four. I'm probably not the most. I'm probably not the best person to ask about big four things. My background with big firms. I was coming through an undergrad program that was in the Portland area. And the goal for the whole program was to pump as many of those suckers into big four firms as possible. That was their metric of success, was how many people they placed into big four firms. And man, I got a little smell of that. And I was like, okay, week long training in Disneyland. I could get behind that. But literally everything else about it to me sounded terrible. And so I did literally the opposite thing. Coming out of school, I went to a wee baby firm while all my friends were like, getting bonuses for passing their CPA test and doing all these cool things, I did the opposite. I went to a little baby firm. I got to work with small businesses. And like, that was what was really exciting to me was just getting to work with a whole bunch of entrepreneurs the other day. But hey, you know what? Some people like going to a warehouse and counting potatoes and doing audit for the first two years instead of ever being home and always being on the road. Maybe that sounds like a good time to you. So I'm not like super informed on. On having a take on it. I will say this. If you've ever run a firm of any size, you know just how much a house of cards that whole thing can feel like with staff leaving and staff coming and going. And it is shocking to me that they can keep those machines running. Like, it just blows my mind with the number of people that have left. And I know there's total dumpster fires everywhere happening in big firms, but the fact that they can keep that train moving down the tracks and keep paying out partners and somehow maintain clients just blows my mind. So it is a big, complex, ugly machine, to be fair, as most big companies are. Like, you run a company beyond a certain size, it's not going to feel personal. It's going to feel like a big old ship to steer. And so, like, big accounting firms are absolutely victim to all those things. But I think in many of the same ways that big, like, really big businesses are. So it just depends on what you're at and what you're looking for. Is there value in a young person coming out of college and doing some time in big four? There may be, like, the training, resources, the ability to invest in those people. It looks totally different at a big firm than it does at a little baby firm. And that might be worthwhile for you. You want to know what I was doing when I was 20 years old, when I wasn't working? Nothing of value. That. Virtually nothing of value. So could my time have been better spent, like, working? And like, what was I ultimately missing out on when I was working in my 20s? Not much. So, I mean, it is what it is. In the same way that I would never enjoy working for any big company. I wouldn't enjoy working for a big firm like that. So I do think there's a good number of those people that watch my content. And I know I've inspired some folks to go out and start their own firm, man. Absolutely consider it, especially if you're on the fence of leaving the profession altogether. Hear from me, hear from the people in these comments, man. There's like, a more sustainable way than being part of that meat grinder. And the way you get confidence to do that sort of thing yourself is to find a bunch of other people that are trying to do that sort of thing themselves. So you're in the right place. But if it's not for you, it's not for you. It definitely wasn't for me, but I wouldn't go as far as to, like, globally dunk on big firms. Yeah, like the timesheet thing, the, the, like the expectations, the billable hours, man, none of that stuff sounds fun, but I'm not gonna sit here and say, yeah, no, you could totally change to this in a big ridiculous firm like that and it would totally work. Like, I don't. I'm just a guy that makes videos on the Internet. Carlos Garcia said improv is the best medicine for everyone. Oh man. There's so many transferable skills and mindset shifts. Improv can help you make in your personal and professional lives. One, Losing the fear of the unknown and just showing up. Two, Improving listening skills and being more present in the moment. Three, the yes and concept. Four. Not being afraid to fail and learning from mistakes. I've done improv twice this week. I'm having so much fun in improv. I shared that I was doing like some one on one classes. Everything in your life I've learned is calculated. And that's what everybody in your life tells you. Think before you speak. Everything that we do is very polished and calculated and improv is the opposite. Like it is just trying to break that part of your brain. Because if you think it just doesn't work, your brain just like stops. And it's terrifying. For me, it's like to get up on stage and think, okay, I'm gonna turn my brain off and whatever's gonna happen is gonna happen, man. Like that'll. That'll put the fear of God in you. But that's why I love it. Like it's. And that's why I'm just having so much fun with it. It's like a drug right now. I will say this. Last Sunday, I've been having problems with my neck. I got like a pinched nerve sort of thing going on and I was doing improv Sunday night. At the time I was a pirate and somebody said I was a pirate with long hair, went to do a hair flip and I just jacked my neck up on a whole other level. So that's complicated things in the last few days, but it's. I wouldn't go as far as to say it's improv's fault. Couple more questions here. Would you rather fully commit to a dedicated practice management platform or try to continue to custom build a solution with integrations? For example, between ClickUp and Content Snare? I just have the feeling you've walked both paths already. Ain't that the truth. So actually in my firm, I had my cast practice on ClickUp. We had just moved the tax practice to Canopy. And I always give the disclaimer like there's not an absolute right answer on the tools that you use. It is very firm, specific. And historically, I've always been an integrations guy. I want best of breed and I'm gonna integrate Those best of breed things. But if you heard yesterday's show, man, that might be going out of style with AI stuff. So the question of would I fully commit to a dedicated PM platform or try to cobble together different things? If I had to decide today, I would be investing in an all in 1pm platform that was leaning into AI and that was going to pull my documents, my projects, my transcripts, all that stuff into the same system that I'm emailing out of. I think that's where we're headed on things. That being said, it's never going to completely replace the value of these sort of surrounding tools. That's just how innovation happens. So ultimately it will not remove these things altogether. But the value of having them in one place has never been more. If you're already sailing in a direction where these things are separate systems, I don't know that we're far enough along to know exactly what the right answer is. Because even if you were 100% decided that the answer is to go to an all in 1pm that will do this stuff. Well, who's gonna do that best right now? Is it one of the incumbent companies, Canopy just ship something great? What's Carbon gonna do? Is it one of the existing companies? Or is it one that isn't even out there yet? Is it one that's going to come of age Nai and ship an AI first solution? Is that ultimately gonna be better? I don't know. So it's like, I think too early to put your finger on. Like, okay, here's the ultimate better solution. Let's make the jump. Now. The only thing to be mindful of right now is building organizational history in a system that won't let you get that information out. And there aren't many of those anymore. The desktop systems were awful about that. You like all of your organizational data in there is all in this proprietary format that they won't give it back to you. But like, if you're in ClickUp or Notion, it's not hard to export any of that stuff. So the only thing that could backfire on you today is if, like you're pouring organizational history into something where you're never gonna be able to get it back out of that. But there aren't many systems like that anymore. Would you recommend content snare for the client to approve a draft payroll is another question. Somebody that sent me a dm. Generally, if you send me a dm, I won't share your name because I don't know if I have permission to so would you use something like content snares? Content snares are request system to approve a draft payroll or keep that kind of approval as simple as possible. Like just an email they said. Difficult to find the balance between controlling the whole process and still ensuring a simple and good client experience. My opinion, I'm going to be really explicit about how any type of requests are managed versus how the human squishy stuff is managed. And for me a request of any type always needs to go through the same platform and that is a machine that is doing the requesting and a machine that is doing the auto follow ups. I think anytime you deviate from that it creates complexity and kind of expectations and gets humans to doing requesting and doing follow ups. So that which I think is something you want to fundamentally avoid. Let the, let the machines do that. The humans will stage the requests. But as soon as you have humans emailing other humans asking for things, I think it kind of sours that advisory relationship and makes those interactions more about just doing the thing rather than adding value. So in my opinion I would use a request system for even something like approving payroll. All right, last one. Matt Bontrager, I'm sorry, says please help. Tax prep communications were such an issue this year, keeping clients in the know. They use tax dome for messaging, but how do you keep it personable? How do you keep it personable but yet scalable so clients don't hate it. So the age old balance of the personal touch versus automated messaging, I guess this kind of ties in with the last question. It is hard and it's going to continue to get harder as the stuff we use automates better and like frankly as suggested emails get even better and better and like will almost reply themselves without any sort of human tinkering. But where I draw the line, that's pretty much the same thing I said in the last answer. I'm okay setting the expectation with clients that any requests will always be done by a machine and anything that's not a request will always be done by a human. There's a risk of creating noise so like frivolous or low value communications. So I'm 5050 on whether status update emails are worthwhile. In some ways they could potentially dilute the other more important emails that I want CL to see if I'm sending them an Update for all 16 steps in a process. The right answer there is probably pretty contextual. If it's a four step process, probably not a big deal. If it's a 20 step process, maybe more problematic. But I Try to limit the machine communication to just when we need requests, when we need stuff from them. And all the auto follow ups, with the exception I guess of like you know, a weekly or monthly newsletter or something like that. Because it is true, like the more noise you pump through the the less attention they're going to pay to those human emails probably. And there needs to be a really clear expectation of what's one and what's the other. So the expectation I'm perfectly happy setting is anytime I need something, a machine's going to be following up and it's going to keep following up and pestering you until you give it to me. Anytime I want to have another conversation that is not around, simply gathering information that's going to come from me. Other smart things I've heard from people. Anytime you're starting or ending a project, do it with a human touch. Seems smart to me. So have a human kick things off. Have a human kind of put a bow on things and finish off a project that makes sense. But that is admittedly a hard thing to balance. Tax Dome, especially like there's such such flexibility in automating around each step of the process. It could become easy to like get into like overkill on the amount of comms that go out. But you just gotta set some sort of expectation with clients, like what are the boundaries for machine generated emails that should be explicit. And the most logical place for me to set those boundaries is just around requests. What do y' all think? If you have other thoughts on that, drop a comment. As always, I appreciate y' all for coming and hanging. Drop questions you got in the comments, shoot me DMs, whatever's the easiest way for you to get that to me. I'll get more questions than I can keep up with day to day, so maybe we'll just dedicate Wednesdays to like a Q and a roundup or something. Personally, I love stuff like this because it hits on a bunch of different things and there's usually a couple ideas I'm able to steal. But I also love like soliciting. What are your thoughts on all these questions? I think I come off sometimes more deterministic than I really am. Like that I'm 100% set on something and generally I'm not. So like a lot of the value to me of sharing these things is hearing what you think and how maybe your lived experience is different and something that I haven't been able to see into. So please do keep sharing that stuff. Bless all your little hearts. Thanks for being here. And what is it today? Wednesday. I'll see you tomorrow. Thursday. Bye.