016 Software Change Management In The Age of AI

Oh, it's the big day. Oh, man. If you are a US tax preparer, there's just no better drug. There's just no better drug than the tax deadline. It is probably the best feeling I've ever felt in my professional life. I don't know that it's a healthy drug. It may come from a place of kind of like if you're released from a hostage situation, that probably feel really good on the other side. But hey, for all the people out there who don't do taxes and they look at the tax people and they're like, whatever does it for you, man. Not my thing. Hey, we should be feeling a lot of gratitude for these people. Cause somebody's gotta do the work. I don't wanna do the work. So to all the tax people out there, thank you, bless your hearts, stick with it. And I hope you do it again next year because we need those people. Okay, today, what are we talking about? We're talking about change management and what is a really weird time for software right now. We're always, you know, considering new software, that sort of thing, particularly for tax people. When you come over a deadline, the brain kind of shifts into how we're going to change our processes and the tools that we use and all that. But how do you buy software right now when AI is like every headline that you see, what's about to change? Should I really, like, commit to something right now if things are going to be wildly different, you know, 12 months from now? It's tough. I got some ideas, kind of a framework that I think works. Let's talk about it today on Jason Daily. So Bill Gates put out a blog post recently. Couple interesting tidbits that kind of get me to a framework for how do you make decisions in a time where everything's changing? A few interesting things about the blog post. First, interesting just the fact that it's titled the Age of AI Has Begun. There's a. We talk a lot about people being opinionated about AI and how. I think that's not helpful. We really just kind of need to buckle up and, and take things day by day. But the AI minimizers will be very quick to point out that AI is not a new thing, and it's not a new thing, but what's fundamentally changed is the usefulness of it and the power of it. And so interesting to me that this blog post is titled the Age of AI Has Begun. But he points out that he says, in my lifetime, I've seen two demonstrations of technology that stuck with me as revolutionary. The first time was 1980, when I was introduced to a graphical user interface that is like a computer display that will give you feedback and, like, your inputs go into there and outputs come out. He says the second big surprise came just last year. He was meeting with OpenAI and they were kind of going through some benchmarking stuff, and it just kind of blew his mind how quickly it had all progressed. So, again, ultimately, we're all guessing about what the future of AI has in store. I will happily defer to people smarter than myself, but interesting that he's drinking the Kool Aid like everybody else. And then a couple specific things I wanted to highlight in here. He says, eventually your main way of controlling a computer will no longer be pointing and clicking or tapping on menus and dialogue boxes. Instead, you'll be able to write out a request in plain English. And I've been thinking a lot about this software, like, post AI, and I guess I'm calling like now and forward post AI. I think it's fundamentally different because, like, this age in software will look like those cars with like 90 buttons on them, and there's like little bits of chrome on all the buttons. And you're like, what do all these buttons do? They've all got little icons. You're not sure what it is. I think that is like, this stage of software that we're in today, because now, and we've seen this with ChatGPT, the ability to semantically say what you want to do, like, that's enough. Natural language processing is good enough now to where that really eliminates the need for all of those menus and sub menus and all of that. And so I think we're going to a version of software that is fundamentally devoid of, like, all of these fiddly controls that we've had in the past. What that looks like, exactly, I don't know. An interesting thing happening now is ChatGPT. They're rolling out plugins so that apps can interface with ChatGPT. And that almost is like they're trying to make ChatGPT the hub of using different bits of software. So if a bit of Software integrates with ChatGPT in a meaningful way, do I still have to go out to use that software? Or is the future of software just things that integrate with that communication hub, whether that's chat or voice or whatever it is pretty like, I mean, huge paradigm shift from what we're used to today. Another thing interesting about this statement, you'll be able to write a request in plain English. I think language barriers are set to completely change. Let's make another Hitchhiker's Guide reference. Hopefully I don't make people mad like I did last time. The universal translator, the Babel Fish. One of the big developments in AI the last couple years has been vastly improved quality of transcription and cost effectiveness of transcription. We've always had auto captions, but they were like comedically bad for a long time, but still better than nothing. Now that stuff is getting like very near human level quality. And so they call these AI models speech to text. That is, they take human speech and convert it into text. But notably that is speech in any language to text, not in any language. Yet English is the most supported. But ultimately what that looks like is the ability for the AI to hear speech and on the other side give you text so you could then read that text or that could also be passed to a speech model to then say that text. And in an age where we're all running around with like AirPods in our ears that have microphones and speakers and all that stuff like that universal translator, like that's a thing. Like that was a wild sort of science fiction thing in the past, but that's absolutely doable and practical now. How does that change the world? That's actually really interesting to me. So software interfaces fundamentally changing Language merry. Language barriers fundamentally changing last. Here's another excerpt from the blog post. He says, in addition, advances in AI will enable the creation of a personal agent. Think of it as a digital personal assistant. It will see your latest emails, know about the meetings you attend, read what you read, and read the things you don't want to bother with. This will improve your work on the tasks you want to do and free you from the ones you don't want to do. I think we it's easy to lose perspective of how limited ChatGPT is right now. The fact it's in this box and it can't remember things and it's not plugged into your other apps. But man, you take that box away and that's like such a big change that I honestly can't even get my mind around it right now. But in the very first ChatGPT video I did on the main YouTube channel, one of the use cases I talked about was as this sort of like personal journal. And especially when it comes to voice, like making a voice memo, a lot of us just process things better vocally. And so what that could look like is maybe you take five minutes at the end of every day just recording a voice memo and that goes into your own sort of personal AI. Now, the hard thing with this right now is GPT context limits. The fact that it can only remember so much at a time. Right now with chat, GPT that's its limits is the equivalent of like 8 minutes of spoken word or so. But with other types of technology, you can build it in long term memory to a GPT experience. And those are called vector embeddings. We may talk more about those someday. But fundamentally, let's say you have this sort of personal journal and it could be written, it could be vocal, it could be anything. You use this personal journal as a way of almost starting to invest in your own personal AI, right? And this thing gets smarter. Maybe it's business stuff, maybe it's personal stuff, maybe it's everything. And so you have this sort of personal assistant that becomes really powerful. Equally interesting is what about the AIs of other people of other shows? Let's say there's a podcast that you love, and that podcast already has this massive database of context from all these podcast interviews that you've learned from and all of that. What does it look like to license that context so that you can pull it into your own personal model? So that when you're querying for something, it pulls in that helpful information, those helpful interviews and frameworks, stuff like that. This show, Jason Dailey, we're building a huge library of context right now. If you think I say smart things, is there a version of this that's helpful for you where it can then be trained on the context from all these shows that we've done? And you could then pull that into your own personal AI to like add some seasoning or external perspective that's not just you talking to your journal. Right. I'm so excited for that. I'm particularly interested about, like what licensing looks like in that case. Cause even through the lens of an accounting firm, internally, let's say you're a manager, you can have your own model, your own context as a manager about how to work with you, about how things at the firm work, about training, about how to use QuickBooks, about how to use the tax software and all these things. And your team can leverage this. Now, it's not a replacement for talking to the manager ultimately, but it is an asset that you can build for yourself that then your team can chat with and save you a ton of time, give them easier access. Because being a Junior is 90% stumbling through things and eventually having the confidence to ask the question that's just Gonna make you look like an idiot. And if you have a low stakes way to do that, and you know that's really one of the beauties of ChatGPT is you don't have this fear of asking the dumb question like that's really valuable. So internally in a firm you could have your own little personal AMI models for this purpose. Externally, obviously if you're a thought leader on a specific topic, you can have things that clients can interact with to ask those dumb questions and that becomes useful. But then even contextually, what about a model that only knows the that client's information? It sees the source documents that it gave you, it sees the tax returns that you filed, it sees the monthly financials. I think context, you know, everything is context that can be gobbled up by AI in one form or another. How we delineate and draw boundaries around those different types of contexts will be really interesting. And that becomes a really viable resource for your clients to be able to ask questions and stuff like that. Ultimately it's never gonna be a complete replacement for having a conversation, but there's a lot of things short of having that one on one conversation that are still valuable, that are still a whole lot better than Google, right? That are way more contextual than what you'll find online because it can see that client's information. Really amazing, powerful stuff. I'll put a link to this Gates blog post in the show notes if you wanna check it out. But where does this get us on? Like that's really cool. That's really exciting, Jason. Like age old thing. I know with my content. That's really cool. Okay, what do I do about it tomorrow? Like what does it look like to make software decisions in this really fast changing time where it's really hard to figure out if this is going to be fundamentally different 12 or 24 months from now? Like, why would I make that change today? And so I think there's a degree of AI adoption right now actually slowing slow software adoption because people don't know what to think. You've got the anti AI crowd saying, nope, you just gotta stay the course. You gotta stick to what you know. You've got all of these buzzword headlines that are saying like, absolutely everything's gonna change. And there's like things that we know will happen, but then a lot of it's just everybody's guessing and we talk about this a lot. The fact that like in the AI age, it's just a matter of being aware of what's happened day over day. You're probably going to have to change. But if you're plugged into like, understanding what's happening, then at least you won't be surprised by that change. You shouldn't be displaced by change because you saw it coming and you could get, you know, more specific about what you do or just more intentional about how you build your business. And the best analogy I can come up with is it's kind of like, like the iPhone refresh cycle. What is the right time to buy an iPhone? With iPhones, at least you generally know the annual cadence where they're announced. And this goes with like all flagship phones. They're generally released on the same sort of annual cadence. But what if that next one has like this killer feature that I absolutely need? Probably a bad example now because it's not really a thing with phones anymore. Cause they're all the same. But if you live in constant fear of that thing that could be released six months from now, that will like, just be fundamentally better than the thing that you have now. How do you ever make a decision? Like you're just perpetually paralyzed by what could be right, which isn't really a way to go through life. Like, you're not actually going to then be like, making meaningful progress. And we are paralyzed by this issue. But you know who else is paralyzed by this issue right now? Oh, buddy. Imagine being a software company. You think you're scared for the accounting profession? Imagine being a software company who sits down and reads that Bill Gates blog post and says, oh, yeah, you don't need user interfaces. Well, crap. And I've, like, I've talked about this. I think there's gonna be like, I mean, throw a percentage on it. But let's say half of the products companies out there will be like, okay, we're gonna lean into AI, make the most of this, and they'll be around for a while. But then the other half of tools that we may be using five years from now, maybe companies that came of age in the age of AI in the same way that when we went from desktop to cloud, when the desktop companies went out to make cloud apps, they kind of sucked, right? Like, you could tell the companies that were SaaS first and they had that expertise, that's where we're at today is just at a different inflection point where we're going to AI. So the companies who made their hay in like the pre AI era, they're like, well, crap. Do we hang on to what we know right now and what we do really well, or do we Throw out the roadmap and try to do something different, something that's going to lean into AI. Honestly today I would rather be an accountant than like muddling through our issues than that software company where the whole product now is like up in the air. I mean we just saw like a couple weeks ago how we're pulling bank transactions off of bank statements with just vanilla GPT. I mean that was Auto Entry. Like that was the core functionality of Auto Entry that was sold to Sage a few years ago. It's not gonna do absolutely everything Auto Entry does, but it's not a great time for software developers. And so they've gotta decide they've got their own version of the iPhone issue. Do we start making a whole bunch of changes right now knowing that very possibly tomorrow OpenAI could release image processing for GPT4, which could come any day. And that fundamentally changes how you approach stuff like OCR and pulling information off of documents and stuff like that. Where do you draw the line on waiting and seeing versus okay, this is the time that we're gonna jump in and do it. And for me the only framework that I can get to that is seems manageable. And it's largely the same for firms as it is for software companies as you kind of have to manufacture your own sort of timeframes and milestones around what you know today. And then ultimately you know that like those milestones are going to be shorter, they're going to be more compressed than they were in the past. Your ability to be agile now will be rewarded more than ever. But you got to just at a certain point like draw a line in the sand and say, okay, we're changing our practice management so system. And with that decision you may be saying and we are not going to change that for you know, 24 months or something like that. For big decisions, it's hard to make big decisions right now. For small decisions like spinning up ChatGPT to solve these fiddly little client specific things that none of your apps are ever going to handle. That's easy. That's great. Let's say you don't record any of your meetings right now. So you don't have like transcriptions of those meetings. That's easy man. That's a like in my mind a no brainer these days. And you go pick up, you know, a new app like Fireflies or Fathom or something like that that will do transcriptions for you. That's an easier decision. You could start it tomorrow, you could turn it off and it's not gonna be the end of the world. But for the bigger stuff, the more kind of infrastructure types of things, I still think you just need to, like, draw a line in the sand and say, okay, today this is the right answer. We're not gonna touch this again for X amount of timeline. And that's it. Because there's so many shiny objects right now that, like, it's okay to be impressed by them and go out and check them out and all that. But the way that you make decisions on software, like, you still have to kind of manufacture your own framework around that so it is logical and not this thing that is constantly in flux, not only for your own sanity, but for that of your team as well. Because most teams generally aren't going to enjoy going through software transitions. But if you can say, like, here's kind of a framework for how we're gonna think about software transitions going forward, here's the cadence on which we're going to review what we're doing, then at least everybody has, like, kind of a common set of rules that they're all working against. Same thing with product companies and, like, development plans. If the rug's being pulled out from under you every single week, that's not only gonna be frustrating for you, it's gonna be frustrating for the entire team. And so the balancing act to me between perpetually sitting on the sideline to just, like, waiting for the best iPhone until the end of time, like, obviously that doesn't work, and being super distracted and wowed by every single thing and changing too much, is something in the middle where you come up with your own organizational rules for how much change, like, you are willing to adopt. And I get that's still kind of squishy. But that's why I think you gotta sort of manufacture your own rules around how you're going to approach that now more than ever, when you're making those decisions on what software am I gonna use, that sort of thing, you are well and truly hitching your wagon to something that should be changing really quickly. So on both sides of the camp there, from the accounting firm user to the product company, it needs to be really clear what the plan is to build AI into that product. And there is virtually no products out there that will not be impacted by AI in some meaningful way. So if you're the product company, like, I get that it's hard to know what's going to happen and all that, but I've also seen a lot of product companies and I've talked with them and they have like zero AI vision. They're just not investing in learning and figuring out what makes sense for their product. So you don't want to look like those people. So at the end of the day, you need to be able to communicate what your plan is to lean into AI. Because when people are making buying decisions, that is one of the top questions that they should be asking right now. Cause the last thing they want to do is go and like pull the rug out on this software tool that they just picked up when there's a great AI enabled version of that that comes out six months from now. Right. So as we're kind of talking with the apps that we're picking up, that I think has become a really important part of the discussion. What does the AI roadmap look like? Like, what are some exciting applications for AI and what you're building? That was like a big part of the reason why I put out that big long list of like, hey, here's your, your practice management management system on AI to kind of light a fire under product companies, but also to get people excited about what's possible with today's tech. Like, it's, you could absolutely go out and build it tomorrow. Not to like understate what a project it would be to build it, but that stuff is ready. Like, there is no question of what's ready, what's secure. Like, that stuff is ready to go today. Okay, Put the old shower caps on. Shower thought. I've heard of various, like framings of this same thing and I love it. And it's like how I'm trying to live my life. The idea is everything that you need to know, someone else already knows. It's in someone else's head. Your job is to get that idea or that concept from their head. So I went through like my, the first decade of my professional career, like doing what I thought was like making really cool stuff, building all this stuff from scratch. Absolutely everything was from scratch. I had no network. Everything was just like me hacking away on this cool stuff in a vacuum. And I think like in the accounting industry, most people largely go through their whole lives doing this. They're not going to conferences, they're not networking with other firms in town, other firms online. And your perspective and viewpoint of the world and how to run a firm is limited to that of just you, your own firm, your partners. And when there's like millions of people out there doing this stuff, it just doesn't make sense. Like there, like there's always more to gain by talking with other People. There's this, I think, inherent arrogance in thinking that you have nothing to gain by talking to other people. It is as if you are number one in the world at all aspects of firm running. You know, to be able to say that you have nothing to gain by going out and talking with other people about what you do, even if it's different firms, even if it's people who are 10 years behind you. In my experience, virtually all of those people have some aspect of firm running that can add seasoning to how you approach decisions. So what does it look like to approach your days through the lens of not just doing the work and using our kind of organizational group think to make decisions, but to optimize how you do that according to all the other ideas that are in other people's heads that you haven't unlocked yet. This was like what blew my mind a couple of years ago and was why I started my accountant community. Realize is I got to the point where I'm like, cool, man. Like, you could waste a whole bunch of time, like, reinventing the wheel and doing that stuff yourself, or you could just go out and find the people who did that thing three years ago, and you can cut years of struggling off your life because the likelihood that you actually have come up with something that's never been done before, that's pretty low. I like to think I'm a creative guy, but at the end of the day, is it, like. Is the thing I'm thinking of, like, truly unique? Is it really totally novel? Probably not. And if that's not the case, then there's somebody out there that can help me. Then there's a shortcut out there that I haven't found yet, and I should go out and look out, look for that shortcut, right? And, like, the. The example I always come back to is a practice management system transition. We went through, like, five years ago, and we've spent probably nine months, like, looking at different systems. And this was like we were migrating, like 35 people do a different system. And we were like, two months into the migration, like, into using it. And so the whole process probably gone on a year at that point. And we were a couple months in, and I was confident we had made the wrong decision. That was just not the right choice. And this project, this practice management system, was everything. It was documents, it was projects, it was time and billing. It touched every single aspect of our firm. And what is the cost of that? A firm that size having sunk a year into that? What is the cost of making the Wrong decision on the thing that runs every aspect of your firm. It is astronomical. I could have spent a year talking with other people about, is this the right thing? Should I do this? Like, for our firm, what looks best? Had I had those conversations, never would have made the decision that we did. It was not just the wasted effort that went into making that transition. It was, oh, crap, what do we do now? We're on the wrong system. Every day we are burning time and energy and frustration on, on the wrong system. But the last thing anybody wants to do is go to a new system now like that, like, everybody is so burnt out by change. That's the last thing in the universe people want to do. And that's like a huge momentous example of go out and steal a bunch of people's ideas before you make a decision. But there's a lot of smaller versions of that that we go through every day too from like that fiddly client specific workflow in the month end close that you're just wasting a bunch of time on every month because you're hacking it together in Excel. Post a tweet, post a comment to this video, like, what is that fiddly little thing where maybe you can't get the gusto integration just right so you have to manually do this calculation. There's a bunch of smaller versions of that out there where the more you put yourself out there and the more you're willing to have those conversations and find the people who are willing to be collaborative, of which there's a bunch of them out there, the more it's ultimately just going to benefit you. Because when you put something out there, and this is why I post online so much in return, I get like the collective experience and seasoning of all these other people's lived experiences that I have not lived. And in that case, who wins, like, who wins the most? I win the most because I see all that, the other people replying and reading the replies, they're learning from all that stuff. So like, what does it look like to go through life optimizing for cracking the noggins of everyone else who's got the good ideas, getting, getting them in your network, being able to have those conversations so that your worldview is way more robust than if you just have that tunnel vision, doing your day to day work in your firm with the same clients you see every day, with the staff that you talk to every day. I love that everything you need to know is already in someone's head. Your job is to get that idea out of their head. Think about it. Hey, Happy tax day. Thanks for coming and hanging today. And I'll see you tomorrow, Sa.