The Intelligent Firm: Setting the foundation for 2027

Trent McLaren: Now, I'm not used to introducing myself for anything, ever, really. I like to do this part. This is the fun part. There's a lot less prep involved when you're doing everything else. But today, I'm actually running through what I would call our thesis, my thesis, on what we think the future of work looks like in the accounting industry, and this is particularly from the Vinyl point of view. Now, I'll explain a bit about that in a sec.

Trent McLaren: For those that don't know me, well, you've seen me on webinars, and I jazz hands say hello to a lot of people all the time, see me at events, hosting parties, all those things. I have worked in the space for about 14, 15 years now. I've worked at a lot of the different software vendors that you all know and use, been to a lot of conferences, traveled many countries, spent more time on planes than most people do in a lifetime.

Trent McLaren: And during one of those trips, I got access to Starlink. Starlink on a 15-hour flight is amazing. Well, it's good and bad. It's amazing because I got so much done. But I had this thought and idea for a project called The Intelligent Firm, and a lot of the basis of this is that it should be a book, or a collection of things. My only issue with this is that things are moving so quickly that how I'm wrapping my head around this is changing a lot, but the overall lessons and rules are in there. So The Intelligent Firm is a book that I'm working on as a project, which basically becomes the spine, or the thesis, for what we're building at Vinyl.

Trent McLaren: If you don't know what Vinyl is, Vinyl is an AI meeting assistant. It records and captures what you say with your clients in meetings. It transforms those notes into actionable things you can do after, things you can share with your team. But I'm not here to talk to you so much about the Vinyl side. It's everything that wraps around that we think plays a future role in what the Intelligent Firm is.

Trent McLaren: Now, this presentation today is giving you the core key themes of the book, condensed into a 40-45 minute session, and once the book becomes available, I expect that to be early next year, possibly, we will see. I'd love to get your feedback, I'd love to hear what you think. All of this comes from my opinion, what I see in the industry, the accounting world, and I think this is where it's headed.

Trent McLaren: For me, personally, I spend too much time in AI, I would say, actually. I've pushed myself so far that I actually enjoy the idea of going outside and appreciating the touching grass part, and if you've been deep in anything AI this year, it can get really overwhelming. You can go into some really dark rabbit holes for 4, 5, 6 hours at a time, just wondering, can I do this? And what I've come to learn is, yeah, I can do this. So I now have a new filter of, I could do this, but I'm not prepared to sit here for 2-3 hours seeing how far I can take this rabbit hole.

Trent McLaren: And I guess out of all of those thinkings and learnings and everything I've picked up, the way that we interact with people, with technology, with process, is fundamentally going to change. And some of you are on different spectrums of that journey. I heard that pretty loud and clear yesterday. Some people said, Trent, I don't even know where to get started, whether I use Teams or Copilot, or Claude, or whichever. There's some people at the start, there's some people at the top end, the top 1% of what accountants are doing in AI. There are some people that have been speaking today and yesterday very much in that top-end spectrum, and we're all here together trying to learn how to get to that part.

Trent McLaren: That is the goal. You don't all need to be top 1%. Ultimately, it's what do you want to do with your firms, your careers, with your teams? That's what, for me, it's all about. What are the outcomes that you're actually trying to hope for, and how does AI help you or hinder you in your approach to doing those things? That's my hope for you.

Trent McLaren: Coming back to why do I think this matters: I think this is the future of work in accounting, and you'll hear why. Not all of the future of work, but components of. Now, I'm going to talk about some things you could do in the belief of what we're trying to do at Vinyl, but there's very much ways you could do this without Vinyl as well. Not everyone will ever become a Vinyl customer, and that's okay.

Trent McLaren: Why now? Why does this matter now? Why are we so focused on this being the thing? Why do we think this is the future of work? AI is at a point where it's actually finally capable of doing the work. A lot of people would argue that Opus 4.5, I think, came out October, September last year, and it was awesome. It was a really big step up. We moved away from just doing pirate poetry to see what you could do in ChatGPT, into actually, hey, can you look at this, and then draft my email, and then update someone in Teams, and then do all of that while I go and get a coffee? That's a reality of where things are going. We're finally capable of doing the work.

Trent McLaren: 2023, we started with a chat window, then we had projects, so we had more context, the idea that systems could remember your projects, and you could get up to speed quicker. You weren't having to re-explain everything every single time. We then got MCPs at the end of last year, and I think the MCP was really that next level up that allowed the co-working component to start doing the work, because now, when my AI wakes up, it's already checked my inbox, my calendar, my CRM, my Slack channel, and by the time I've poured my first coffee at 8am, it's already triaged my inbox, drafted all my replies, and told me the things I actually should care about.

Trent McLaren: And is it perfect? No. But is it extremely helpful? Absolutely. And for all of you, I think your inbox is actually the easiest way to get started with anything in AI, just letting your AI go and tackle that. But this is where we are. Co-workers has very much been the theme of 2026, and I'm going to uncover a bit more about what I think 2027 and beyond kind of looks like.

Trent McLaren: This is all about setting you up for your infrastructure. How do you let this operate at a firm level, top-down, all the way through? Some of that starts with you, some of that is going to start at the team level, and then work its way through. Some of you are already doing some of this as well. I can see in the chat some really smart people that know a lot about this, maybe even more than I do as well.

Trent McLaren: Now, things that you're still having problems with: capacity still sucks, you don't have enough good people, you maybe are or are not offshoring and outsourcing, and that is or isn't working for you. Advisory is something that you've all thought about doing, and Ryan Richardson and I talked yesterday about the confidence gap as the thing that stops people being able to do that. It's not necessarily anything other than your own belief on whether you can do it, approach it, have it with a customer, and then have some longevity in those services.

Trent McLaren: More importantly, you're bleeding knowledge. You're losing knowledge every time a team member leaves. That history goes with that person. It's locked in their brain. You're not doing a good enough job capturing all of the context that exists inside someone's head. And that's a big part of what I'm going to go through for the rest of this session.

Trent McLaren: Your clients' expectations are also changing, right? There's more pressure from government, from clients. Clients always want things yesterday, they never want to pay you on time, but they always want something when they want it. They want you to sign their accountant's letters so they can go buy that car that's definitely for business purposes.

Trent McLaren: And second to that, you've got this tool sprawl. You've been buying every shiny toy at every expo. I know, I've seen it, I've heard you all talk about it. We've bought every tool under the sun, and there's two problems to that. You've got too many apps, you're not using all of them properly, you're at about 20-50% on usage of all of them. And the other side of that is all your tools are now starting to compete with each other, so they're all starting to build in, and you're trying to figure out, well, which one do I want to own that part of the workflow? It's a really interesting time.

Trent McLaren: And then there's the AI pressure. You come into summits like today. Some people will tell me, Trent, I just feel like I'm behind, like I'm nowhere near where I should be. And I don't want you to feel that way, that's not what this is all about. But I appreciate there is stress and pressure of, I'm not doing enough, I've not kept up with what's going on. None of that's real, by the way, that is just what is in your mind. These are the things that you're telling yourself, and at the end of the day, you're the only one that gets to create and control the stress and pressure around this component. So if you feel like you're behind, I promise you, you're not behind. You're exactly where you need to be. And if you want to be somewhere else, then that's on you too. You get to make that choice as to where you go next.

Trent McLaren: Overall, there's slightly different symptoms, but ultimately it all comes from the same root cause. Now, my version here of the Intelligent Firm is that it's a firm that actually hears itself, and I'll explain more now. That comes in 5 ways. We want a system, platform, infrastructure that captures every conversation, every channel. Understands what was said. Is it getting routed to the right place? Does it remember? Does it have the context? Is there a brain, a memory, a system log of everything that's ever happened that can then take action on the other side?

Trent McLaren: Now, this isn't about you buying 10 AI tools and going, hey, we're a really intelligent firm now. This is about connecting all of the dots. And this is not the first time you've had to do this. You are already more likely to buy a tool that integrates with your existing system, right? You're using a practice management system. Obviously, we want tools that integrate with our practice management system. We do not want to operate in standalone environments. Nothing has changed here. Anytime you're buying your AI tools and solutions, it needs to work in and around, and most good AI tools will, so that's not necessarily the thing that I'm most worried about.

Trent McLaren: There's 4 key themes and concepts I want to run you through right now. The idea of the voice operating system, and this plays a big role in where I see us going as a company and organization on the Vinyl side of things as well. Right now, when you use AI, you ask it, and it answers. You open the chat window, you type in the thing, you hit enter, it gives you a reply. If you're a little bit more advanced, you might be speaking to it. You speak into the window, you hit enter, it answers, and it comes back.

Trent McLaren: But the idea of a voice operating system, and there's hints of this already happening now, is that it will listen to what's going on. Now, listening could be audio, but it could be reading your email. Let's just consider that it's capturing and analyzing every bit of raw information you've had. It hears the room, it understands it, and it takes action, because it can understand the behaviors and the patterns of what is happening in your firm based on everything that you've ever done prior to that. And it's doing that before you realize you need it. Every meeting, every call, every conversation, it all becomes an instruction for the thing that needs to happen next.

Trent McLaren: Now, the second key concept here is the idea of firm intelligence. You already have firm intelligence. You have manuals, processes on how to do the work, but I want you to split that in two when we think firm intelligence. We've got the client intelligence: the business's goals, what they told us, what we promised. Some of these things, if you're doing it well, live in your SharePoints, your documents, your practice management. Then there's the team intelligence. Everything about your team, how they work, who knows what. Processes and decisions, what we've learned, and whether you're doing that feedback loop and sharing it with each other.

Trent McLaren: These are the two things that you should be documenting and capturing as often as you can. And these are things that are actually really easy to now start doing, if you've got the right system and infrastructure in place. Most of this lives in conversations, some of this lives in your documentation, so that's your policies, your procedures, your knowledge base. A lot of people don't have that. This is where the bleed-out starts to happen. Again, a team member leaves, they take all the good stuff with them, and we are stuck trying to replace a really good hire who had 5, 10 years' worth of knowledge in the first place.

Trent McLaren: My third concept here is the idea of a second brain. Now, this might be the first time you've heard of it. This is not something that I have coined or curated. Much smarter people have been talking about the idea of the second brain, but if I introduce it for you now: the biggest advantage you can get in AI is that you should start to build out your own version of a second brain. This is all the context, the repository of all the information, that your AI can wrap around quickly to get the answers that it needs.

Trent McLaren: On the left-hand side is the first brain. It's everything that's in your head, right? Bob mentioned a property deal, we promised something on that call, but I can't remember what it was, or it's stuck in Sarah's inbox, or there's a sticky note somewhere that said, oh, don't forget to do X by X. These are all the things that happen in the first brain. In the second brain, it's documented. We've got it by the client: they told us they wanted to buy a property, we promised to give the advice, and then something changed. The bank wants the numbers by Friday. All of that context needs to be captured, and that gets built out in your second brain.

Trent McLaren: Now, this is not a replacement for your people. This is about trying to suck all that information in, so that you all have shared knowledge of what's going on at any point in time, and your AI has access to that at all times as well. So you've heard kind of my first stage of how did we get here, what is our current state of play, and what are some of these key themes that we're going to break down a little bit further. I promise there's some more practical visuals for you to take away from this. I want to make sure you've got a good action plan on that as well.

Trent McLaren: Now, for the last 40 years, all the tech has been focused on the accounting software, the numbers, the ledgers, right? Go back to the 80s, we had desktop ledgers, spreadsheets. We've been focused on the numbers for so long, because actually trying to capture everything else was really hard. We didn't touch documents, emails. We were manually doing a few things, but nothing that could do that automatically, because we weren't at the right time. But 2026, 25, arguably last year, that started to change. And what we're seeing over here on the right in the 2020s is that, yeah, we are now starting to get agents, agentic AI, starting to take action on all these things. The ledger, the bills, the variances, the lodgments.

Trent McLaren: There are some sessions today where you'll see AI actually handling all of this end-to-end. Perfectly? No. But as a majority, yes. With reviews and approvals, absolutely. But some of it autonomously, that a human never touches? Also, yes. And that is an interesting place for us to be.

Trent McLaren: So when I think about everything else, we've been scattered, right? Inbox, SMSs, that one client that likes to text that one person with the receipts and everything else. The voicemails, the team chats, the water cooler moments, the good ideas that come up in the office. That then gets pulled together by hand, right? We started manually logging things in practice management, and then slowly we started to see practice management bring those things in. Documents started coming in, email started coming in. Arguably, I think phone calls haven't quite made their way in yet, but it's starting to, which is awesome.

Trent McLaren: And in 2026, these are the things that need to start to get captured and synced and organized into your second brain. Every email, every phone call, your meetings, your SMSs, your documents. AI does a wonderful job of capturing and organizing really messy context, and for me, and our whole Vinyl thesis at the moment, these are the things we think are gonna help make the biggest difference in your work life, and also help with the future of work.

Trent McLaren: Alrighty, so what does that look like? How do we make sure that for every conversation in comes the right action out? I mentioned some of these pains before, but if you've not synced all your inboxes and emails, if you can't see what's going on for a client in your team's inbox, that's in Mary's mail, I need to wait till she gets back, I can't access that, you're in trouble. A partner goes on holiday for two weeks, you're about to head into December, January, where a lot of people aren't coming back till after Australia Day, and there's a few weeks there where everyone else is like, well, we're gonna have to wait till they get back. I can't move that forward until this person gives me the missing part of the information that I don't know.

Trent McLaren: These are all the things that go wrong, okay? It's stuck in areas that you can't solve for. A text message you get late at night that you have to remember the next day, oh yeah, I needed to do something with that thing. This is the issue that we're coming through. All of the intelligence exists. Everything that I'm talking about exists already, but we're not doing enough to capture it and bring it through the other side.

Trent McLaren: This is how I would see this on the other side. So when your client's email, phone calls, meetings, SMS, website, voice notes capture that conversation, your AI then needs to be able to triage, understand, classify it, draft and suggest, and give that to the right person, the right system, into your brain, which goes to your team, or to the partners, depending on that, and within that, we need drafted replies. This is the ideal system that we want to start to move towards, so that everything that comes in that we are systemizing has a way to be reviewed, captured, triaged to the right person, to the right system, and it then stores a memory in your firm's brain.

Trent McLaren: And on this side of it, this is actually a big part of where I see things when I look at our roadmap and what we're doing at Vinyl. Yes, we're a notetaker, but absolutely, we are going after every bit of context available to help triage it into where it needs to go.

Trent McLaren: Another way to look at this is to think about, well, what's the client, and how does this flow through your systems? The client goes through your app, so again, this could be Vinyl, could be a practice management system, could be a few different things: SMS, email, meetings, that one stream. It goes through the memory, it says, okay, I heard this, let me check the context of the client, okay, cool. I've seen the memory log of what's going on with the client, I have a skill, I know how to tackle a request for a certain bit of information, because you've built those skills within your systems. That then triages it to the right person, so that skill becomes a routine. A routine could be a daily routine, it could be an hourly routine. Your AI is constantly checking for a thing in your inbox, or a document, or a channel, or whatever it may be, and then it is flagged to the right person at the right time.

Trent McLaren: We're suggesting that all of this could happen within the minute that it occurred. So if the SMS happened at 7.42, we've captured it. The memory's reviewed it. It's been triaged to the right person. There's a routine to make sure that it was picked up and not missed in the first place, as a task in practice management. A reply was drafted, and it was flagged to that person a minute later. Now, take that out. This may happen the next day, but it may not happen for a week or two.

Trent McLaren: If I look at any data on small businesses and why they leave their accountant or bookkeeper, and take this with a grain of salt, I know you're all very busy, but a lot of the time it says it's hard to get hold of them. It's hard to get back in touch. I always want to call them, and I never hear back from them. I think your biggest opportunity with AI is actually increasing your ability to deliver a service to your clients at scale that was never possible, because we didn't have that team member, the admin, client success manager, account manager roles, that can pick up the phone, that can explain what's going on. I think your biggest advantage with AI is that you can now capture all the context and train it in a way that it can actually draft and get these things ready, and flag it to the right person so that you can then make a decision on the other side.

Trent McLaren: So when the client asks, can you send me the numbers for the bank by Friday, it's for the building, the AI has read it, it's pulled the attachments, it's pulled it from your firm's brain, it's drafted it for that person to then review it. That client then hears back. Could be same day, could be next day, but it's happening on a faster scope, because you're letting the AI do all of the grunt work of working through the context, the opportunity, and what needs to happen next.

Trent McLaren: Okay. Where is this all headed? Right now, it's all about capture and ask. I remember a year or two ago, we were talking about the idea that agents were gonna dominate the media, the headlines. Everything we were talking about was agent, agent, agent. And this year, I would say that has happened, but I think co-work has probably been more so the thing dominating. We're doing things together with the AI. 2027, for me, will be more so about routines, and agents starting to do the work on routines, on schedules, on your behalf, with or without your review. There are some things that happen in my agent workflows that I do not need to review anymore. I'm happy and comfortable and confident that it's doing what it needs to do.

Trent McLaren: This is what will start to happen in 2027. You already build workflows, but you will start to build agent workflows that have consequence if they get it wrong, and that will start to take over 2027. I think this is the thing we'll be talking about until the cows come home. Eventually, that will evolve into agents taking over more of the work. And there will be a shift in, well, if the agent's doing that, then what is this person now doing? And I'm going to talk about that in a sec.

Trent McLaren: And eventually, as you continue to capture your context, your documentation, which will just start to happen automatically as you're speaking, as you email, as you file it, it's in a system that the AI can read, and it's going through your processes. You can see how this evolves. You can see where this is all going. This is my practical view of trying to explain to someone, what do you mean AI is gonna take over and start doing all the stuff? It's going to take over the stuff we actually don't want to do. Nobody likes doing all the admin that I'm talking about. And the other side of it is your client and you should hopefully have a better relationship, better response times, everyone's happier, everyone's moving in a much better direction. That's my prediction.

Trent McLaren: So what stays human? It's the judgment, the relationships, it's the signature, it's your sign-off on the other side. Everything we're talking about is the AI doing the grunt work to get the answer on the other side. You still need to understand it, your team still need to understand what's happened, they're still going to want to see everything the agent did, almost like you're reviewing the junior's work. Our agents, I think, are now becoming those junior admin layers that we need to be reviewing on the other side.

Trent McLaren: I put together a service blueprint, so this is a good way to think about what our workflow looks like right now, without AI, without agents in place taking over the work. So, the client gets in touch by email, the partner or accountant reads it when they can, and that lives in your inbox, okay? The client's waiting for a reply. The service blueprint is looking at all the different layers, right? From the client's point of view, the partner's point of view, the team's, and the underpinning systems that support that part of your service blueprint. So the client waits for the reply, the partner might forward it on to the team and say, hey, can you look after this for me? They ask the client again, hey, can I get some more information about this? The client sends the documents, the team's chasing the documents, and this is all then underpinned by practice management, emails, ledgers, documents, and so forth.

Trent McLaren: You know this flow. If anything, I could probably put more chasing steps in here. This looks a little bit too simple and straightforward. I'm sure it never goes to plan like this. But at the moment, your team, your people, carry the work between the systems. Sometimes an integration does the handoff, but otherwise your people carry the systems, and at every handoff, we're waiting for things to happen at different points.

Trent McLaren: Now, when we introduce agents on this layer, you can see that those first couple of steps actually remove people from those steps. Things start to happen more efficiently and faster. The client says something on a call, or a text, or an email. The agent captures it, classifies it. We're seeing it do this now with our bank feeds, our transactions and coding. Again, not saying it's perfect, but it's doing an okay job. There's some really cool things going on there. It understands it, right? You'd see this in live chat systems. If you've used a live chat system that's plugged into AI, set up well, you can actually get a refund through, I don't know, it could be Uber Eats, could be taxis.

Trent McLaren: I got a refund for a taxi through chatting to a support agent not too long ago, and I was pretty impressed. I was like, hey, that's cool, you understood my point, even though I could have been lying, but whatever. You understood my point, you gave me the refund, I didn't speak to a single person, it was done within that single interaction and didn't bother a human in the first place. Now, look, there's guardrails, there's gaps, there's things we want to do to protect revenue leakage, but the technology is here, and it's actively working, and there's no reason it can't be working for you. There's no reason you don't have a live chat AI agent on your website that's helping to tackle questions and queries from clients that are thinking about working with you. You could be doing this today.

Trent McLaren: I digress. The client gets in touch, the AI understands it, it's actually drafted and sent a reply back already, saying, hey, is this what you mean? Just clarifying context. Okay, cool, this is what I have available, is that helpful? Again, you'll set the rules on this. It might ask a more clarifying question, the client uploads what they're asking for. People now start to handle the exceptions. So if we can get AI to handle 70, 80% of these things coming through, then there's 20-30% of exceptions that I expect your people now start to go through, until these things get better and better. You review the draft, you approve it and send it, and it has the advice on the conversation.

Trent McLaren: Should you be charging any less for any of the things that I'm saying here? Absolutely not. Your clients don't pay you for the amount of time you spend doing things. That has never been the case. You're gonna need to think about that at some point. If you are still billing on time and jobs and everything else, I get it, it works, you've been doing it a long time. This is going to change some of that. You're not billing on time, you're billing on outputs. Token usage. We talked about token usage yesterday as a thing that you're going to need to start thinking about incorporating into your margins.

Trent McLaren: You can see the systems underneath all of this, right? I talked about the idea of Vinyl trying to help capture everything. This feeds your firm's brain. Your agents are connected to your practice management, your ledger, your document management, the routines on schedule. There's 3 parts to how I want you to think about this. Every conversation needs to get captured. It needs to sit in your brain. Your brain might be the practice management system. It could be your document management system, so SharePoint, Count Kit, SuiteFiles. There's so many tools I could put on this slide, there's no particular favorites or anything along those lines, don't take me out of context here, please, anyone.

Trent McLaren: You could be using tools like Notion. We live and breathe in Notion. I never log into it, but I do everything through Claude and say, hey, create the documentation, so that when I go to do things, you know to check Notion, to check the update, to check what's going on, to go and take that next step. You need to pick your AI, and you're gonna need to double down on that. For a lot of you, I think Copilot feels like the most natural fit. Unfortunately, I don't feel like it's been as progressive as the things that we're all loving about Claude and ChatGPT, but I do think Copilot will get there at some point. So if you need to live and breathe and stay in Copilot, do that, play there. I think Claude and ChatGPT are a little bit further along in some of these things, but as we know, things always change.

Trent McLaren: That's been the same for the accounting tech ecosystem as well, but this is how you start to figure out, where is my firm's brain? And all of your practice management systems want this to be the case, right? So when I take a top-down view, and I look at this industry, and I look at what everyone's building towards and where they're headed, your practice management wants to own everything at the top. They want all the context, they want your AIs to go through it all, so that they play a fundamental role in the life of your firm, okay? That's a good way to think about it. They're all gunning for that position.

Trent McLaren: And they're all going to try and bring that AI in as well to try and help you with that. So if you're excited by what your practice management system is doing, I would expect that a lot of the things I've talked about today are things you should be seeing in those systems. And if you're not, and if you're not happy with that, then that's something also for your consideration, as to whether you think you're in the right place to set up your AI infrastructure into the future. As I said before, my thesis and what we are building at Vinyl is that we want to be the system that connects all of the context into your practice management systems, into your documentation, into your AI.

Trent McLaren: Some of you might look at this and say, hey, I might want to build some of this myself. And I actually am trying to build as much as I can. I'm avoiding buying a new subscription like the plague at the moment. I'm trying to kill my CRM, I've killed my AI inbox triage system. Every time I can, I'm ripping and replacing it with something that we can own and control. Our own CRM that's cross-referenced with our product analytics, that looks at our billing systems, where the AI can then say, hey, here's the 10 people you should follow up with today, based on how they're using your products, based on the email conversation, the client sentiment. All of that stuff gets sucked up into that brain, and the AI can then decide how to take action. Some really awesome, scalable things that that unlocks in the process.

Trent McLaren: All of you on this call, there's a good chunk of you that won't want to do that, and I get it. Find your systems, chat to your practice management systems, chat to your document management systems, understand what role they are going to play in your AI infrastructure journey. Are they going to let you wrap Copilot, GPT, Claude around it, so that you can start to run your agents and schedules and everything else? And are they going to continue to play nicely with the rest of your ecosystem? As I said, there's a bit of a war going on in the ecosystem at the moment. When I say war, we're all friendly, polite, we share beers, we still chat, but we're all building each other's products, and that becomes an issue for you, because you're like, well, I don't want them to do that, but I guess that's the thing they have to do.

Trent McLaren: Best of breed still plays a role in terms of your tech stack. You may want to go with an all-in-one. I think the same rule applies. The all-in-ones will still never be the best at what they can do, but they may just be able to do enough to keep you happy with how it's working in your process. More importantly, does it wrap around your AI? Can you start to build in the different parts that you would like to do as well?

Trent McLaren: If you want to build it, though, you can. I am not a developer, an engineer. I know enough to be extremely dangerous. It's not something I'd recommend for everyone, but if you do have the stomach for it, and you're curious, absolutely, you can try to build some of these things. But I would just challenge you on what should you build versus what do you buy, what makes the most sense for you in that sense. The moment you start building, you do have to maintain it, you do have to train team members, you do have to be responsible for it, and you still have billable timesheets and things that you're probably trying to do for your clients. So just decide what role you want to play in that step.

Trent McLaren: Now, I'm coming into the end of my session, but your best team members are probably already using AI. A good thing for you to do to start to make this a reality: what are they using? What are they using it for? What data's going into them? What would you need from an approved alternative? I recommend doing this from the top down. Tech strategy. You have to decide this at the top. You own the decisions at the top as a leadership team. You need to have a leadership team, you need to have a vision for this. Start at the top. Make sure your environment is AI-ready: practice management, documents, emails, your AI. And then, within that environment that you're happy with, that's safe, that's secure, encourage your team, please, experiment, play.

Trent McLaren: Find out what works and doesn't work. You need to do a lot of trial and error to figure out what's going to work for you, because that's all we're all doing. When someone says they're really good at AI, they've put more reps in than you did. They spent more time figuring out what worked for them and what did not work for them.

Trent McLaren: If you don't feel like you're good at AI, it's because you're not spending enough time trying to figure all these things out. So I'm encouraging you: trial, error, ask your team. What does that look like? It's Friday afternoon for 3 hours, 2 hours: hey, here's the task, you're in Claude, you're in our systems, I want you to try and automate this job end-to-end. See where you get stuck, see what works, learn, share, document as much as you can. And it's actually really easy to document. You just ask your AI, hey, we just did this thing, can you document what did and did not work, and put that into Notion for me? It's done.

Trent McLaren: It's a top-down strategy. You've got to set the strategy at the top, make sure everything in your infrastructure is connected, open it up to everyone, encourage your team to play. Now, I've heard a lot about, is this safe? I think the wrong question here is, is it safe; the question is, is it set up safely? You're gonna hear this more over the day, but business plans, never personal plans. Read it, then approve it, until you build confidence in it. And train your team. Get your team on board at the same time.

Trent McLaren: So, is it safe? Look, I believe it absolutely is safe if you've set it up safely. That's a better question to ask. AI is not going anywhere. There are bigger organizations doing more complex, hectic stuff with their data, whether they should or shouldn't be, but in an approved environment, than what you all are, I promise you. There are more people doing more risky things with AI safely than what you all are. My encouragement for you is to continue to push in that direction.

Trent McLaren: So, what are you gonna do on Monday? Pick your tools, go hard on the ones you want to use, connect it to your systems, give it a routine, get it to try and do something on a regular basis for you, to get things started. And experiment, trial and error, and more importantly, you need to start capturing everything: every conversation, every email, every document, everything you tried to do and didn't do. Capture it and expose it to your AI. And then the boring stuff: update your privacy policies, update your engagement letters, let your clients know that this is a new world that you are stepping into, and bring your team on the journey.

Trent McLaren: So for me, my final point here is that your gap isn't the knowledge anymore. We're capturing all the knowledge. It's the decision, it's the judgment, it's the relationship that we are amplified on top of. Draw that line, form your plan, and then make a start.

Trent McLaren: Ladies and gents, that is my session today. It's the first time I've done a session like this in a long time. Let me know what you think. If you think I'm crazy, tell me. If you think that resonated with you, if you think this is a direction that you can go into, please let me know.