A Practical View on AI in Accounting with Julian White

Hello, and welcome to the Big Four Transparency podcast. I'm joined today by Julian White, the CEO and founder of Zentia AI, as well as someone else building some really cool things in the accounting space. So thanks for joining me, Julian. Thanks for having me, Dom. I'm excited. Yeah, so to be clear, you're not specifically just building in the accounting space, but we've had some conversations and that seems to have been a little bit of like a de facto area of specialization for some of these AI implementations. Would that be correct? Yeah. So I would say there's a couple of verticals in targeting initially. Accounting kind of made sense. I have some experience there. There's definitely a need for AI solutions and accounting. So it kind of makes sense to start there. And so far, you know, the feedback from the market has been really good. Okay. Excellent. And for someone who's kind of on the other side of some of those conversations, more as like an employee, when firms and businesses currently are talking about AI implementations that are ready to go today, what are some practical use cases, let's say maybe even more specifically for accounting firms, but what are some practical use cases of AI in its current form today? Because again, I hear all these big promises and, you know, even some fear mongering of people being like, oh, it's going to completely replace us. But that's just really not where things are at as of right now. So what are, yeah, what are some of the best use cases for? Yeah, I don't think AI is going to replace all the accountants and bookkeepers. I think there will be a big, big shakeup. And I think, you know, I think the top 50 people, your top 50% of people are going to remain the bottom 50% are probably going to be doing something else in the next couple of years. And what will really kind of decide that is, you know, the people using AI, those are going to be in the top 50% for sure. So I think the first thing people should do really is just understanding the basics of AI is a great starting point. Keeping up with current tools, looking at what's available. And this is really tough because things are changing so quickly. You know, there's a couple of daily emails I get, and at the bottom of every email, there's new technology, new products that are launching, and there's a dozen every day. So really tough to keep up with this. But you kind of have to keep your eye on this just to see what's out there, what's coming that can really make a difference for you. And I think that's kind of the biggest thing. And then obviously using AI wherever you can, just to improve your efficiency. And however that is, if you're using chat GPT for writing, analyzing financial statements, whatever that is, but you have to be using it and looking at where else you can expand to make a difference. I think trying to make it part of your daily practice or just incorporate into your workflow is very valuable. For me, ideation, I think it's extremely good at that. Naming, you can say like, hey, I'm a big fan of alliterations and I like this type of naming structure and I'll spit out 50 options for you, which even if that doesn't end up being the final option, it'll get you going down some different avenues that might really help you out with where it's at. I use it quite a bit creatively, but I have personally, working in FP&A, really struggled to make a use case out of it professionally beyond just saying like, hey, can you point me to the best resources where I can find these metrics or an update to where these metrics are at in 2024? So yeah, knowing that I've kind of struggled using this professionally quite as much, I'm very curious to know like, yeah, in accounting firms, what are some of these like practical use cases? Yeah, so I mean, to be honest, I think it's easier for the accounting firm than the employees. Because when you ask me to think about ways that employees could benefit, that's a lot harder. Firms, it's a lot easier. I think there's some items that are just low-hanging fruit that can make a big difference. The initial item we had that was really good was client onboarding. So we heard from a lot of accountants that client onboarding was a huge pain, typically took four to six weeks to get a new client onboard, a lot of back and forth, a lot of admin time, a lot of chasing documents. And so what we worked on was a chatbot where you could actually onboard through the chatbot. So the chatbot would ask a lot of questions and it would basically build up a list of requirements. And then at the end, it would give you a form saying, okay, this is what we need from you. There's a checklist of documents you need to upload. It would also remember the conversation, remember the state. So not expecting to do all of this in one visit. So the client can come back and add the documents after they have to go and find things. But the key thing here is they're able to do it faster. It's really clear what's required. And they're not having to call the accounting firm and ask questions about what's required. And in terms of an ROI on something like that, there's an accounting firm I worked with that had eight admins and implementing this, they were basically able to drop down to seven admins and repurpose one person. So there's a 50k saving roughly in salary by implementing that. And having some admin time that you can reallocate within a firm, I think is so valuable. That's a topic I've covered a lot where it's like you have a lot of partners who need to be focused on higher leverage tasks, who keep doing these things that they really should not be doing. And that goes down the tree all the way down to analyst level people as well, where when I was an analyst, I would have killed to have some admin time that I could use, that I could allocate to someone who had additional capacity. So that's, I think, something that's important for people to keep in mind in these conversations is like, I think a lot of places in the accounting industry are so understaffed that if we say, yeah, you can cut from needing eight full time equivalents to five full time equivalents on this thing, that might not necessarily mean three people losing their jobs. That very hopefully means, you know, people going home at five instead of 630, right? So yeah, and those admins, you know, they don't want to be doing some of that work, right? Chasing people up, the things that are missing, you know, they'd much rather be doing something more interesting, adding value to the business. And so what is it about like the AI in this process that makes it so effective? Because I spoke with, I think it was Mike Libby recently, like a Toronto firm owner who talked about they got a custom software built for their firm that was, I think, not AI enabled to my understanding, but was basically just this huge decision tree, right? For client onboarding. So if they respond this, prompt them for document XYZ. If they respond this, prompt them for this other thing. So where is it that the AI like makes that process like much, much better? Yeah. And so I would say this, we still have that decision tree in there. So we kind of have to map out the process, set the rules. So it is kind of like a big flow chart that we put in the AI. But the beauty of the AI is just the ability to ask questions as they're going along, right? So if somebody is uploading documents, or there's a question they're not sure about, then they can actually ask the AI and get the answer without having to call in. And that's the real goal here. We're trying to let the AI handle this whole process, so that a human doesn't have to get involved. Right. And I guess like, you could probably have the AI interpret the documents being uploaded to, right? And say like, no, that's the wrong thing, or that's the wrong format. Yeah, that's the other thing as well. Actually checking that the documents are legitimate and correct can save a lot of time. Interesting. Because I do remember clients had like a really hard time getting the right bank documents that was like a specific issue. Yeah. It's like, send us your profit and loss statement, and they'd send you 84 pages of every single transaction with no profit and loss summary. And it's like, okay, cool. Thanks. Okay, interesting. So client onboarding, that makes sense to me. That seems like, yeah, that seems like a really good starting point for a lot of people. And I do know that like, yeah, a properly onboarded client just from a workflow perspective as well, where there's less stopping and starting less like context switching needed in your work is just absolutely crucial. So good client onboarding can actually trickle into the efficiency of the entire rest of the preparation process for what you're doing. Are there any other kind of like top use cases that you've seen? Well, the first question I ask most accounting firms, how are you taking money? Because there's a lot of accounting firms that are still accepting credit card payments. And that just seems to be a standard in the industry. And so I'll give an example. There was one accounting firm that we switched from, I think they were taking like 25% of their business through credit cards. And that's also people calling in right there to call in, give the credit card details, but the fees are super high, 2.5%. And so we changed that to a pad system, where we can actually send a button to pay the invoice in the email with the invoice. So makes it super simple for the client, they just basically approve the payment, right in the email, and the money comes out of their account seven days later. So the fees on that are outrageously cheap, it's you're talking like 40 cents a transaction as opposed to 2.5%. So there's a huge cost saving there. So this accounting firm I worked with was able to save just under 45k, that's how much it was spending in credit card fees, which is a huge amount of money. And how does that work? Like what's going on behind the scenes there versus just kind of like any old payment processor? So we use a third party to connect the client's bank account and pull that money, basically transferred it into the accounting firm's account. So a lot of this as well, it's good to point out, there's a lot of systems that already exist. We're not having to build stuff, new, off the shelf. There's a lot of third party products out there that we can just put together. It's kind of like a big jigsaw, put them together with a bit of integration. I would say that is one of the lowest hanging fruit. Yeah, that's been really interesting in like conversation with you is I'm like, oh, like, yeah, you're doing like an AI business. What's it like coding all these new AI things, right? And the reality of it is a lot of it exists. A lot of it is having someone who actually understands the use cases, the limitations and whatever, and who can then go and understand your business processes and know where to plug it in. And then sort of like the last mile delivery problem, right, where like you can do the last little bit if it's needed to kind of like plug this in properly. But it is interesting to hear that it's not like, you know, you're not just like from scratch writing all this code for everything, which makes it a lot more approachable for people. Where it's like, there probably is a solution out there, you probably just don't know what it is or know how to apply, right? Yeah, so a lot of the work is integration. And, you know, for the payments example, that isn't even AI, you know, that's basically taking some third party services and adding automation. So automation is another key part of improving business efficiency, kind of often gets overlooked, but it's definitely right up there with AI. Yeah, yeah. And in a lot of roles I've been or have been approached for, like a lot of people actually really like CPAs for like automation type roles or like business process improvement and stuff like that. So that is something like understanding where AI can fall into automation, because a lot of us in more like industry type roles, get put in charge of this type of thing. Again, not necessarily building new products, it's just building new processes and improving the flow of things. So I think that's like a very valuable thing to understand. And then so Q&A chatbots is something we talked about as well. And I mean, those have been a thing forever. So I'm kind of curious what makes these unique with like the technologies that we have today. So yeah, Q&A chatbots have been around for a while and you can train it on your data. For an accounting firm that has like a blog that they're putting a lot of content into, it's great because there's a huge amount of content for the chatbot to draw upon. But that was kind of the V1. I think the V2 is really being smart enough during those chats to know when to upsell services. So if somebody is asking questions about, what are the advantages of having a family trust, the chatbot should know, okay, this is new business. This is a new service. I should probably lead the client towards booking an appointment with the accountant and then let the accountant know, okay, somebody needs to chat about this. And that whole scheduling can be done automatically. It can book the appointment via Calendly right in the chat as well. So that's kind of the V2. And then the V3, I think we're not sure what it is yet, but it's going to be more kind of agent based with the ability to do a lot more during the chat. Right. So I think that's where chatbots will really kind of get into their own when they take on the role of an agent and can actually do stuff during the chat. Right. Like it could probably like kick off an onboarding motion and send out invoicing to be signed and stuff like that. Exactly. Yeah. So it should be able to just bring a client on, get them into the onboarding process, create an invoice, send it, you know, all those things in the onboarding process. Interesting. And then so from like an employee perspective, right? Like if you want to be ahead of the curve with AI coming, if you want to be able to be the one who's the champion at your firm, because I mean, that's a great look, right? Like being the being the manager, being the senior manager saying, hey, have we considered this client onboarding? And that ends up being very, very ROI positive. Like what are some resources that you can do to or that you can use to kind of like, A, just make sure you have a good foundational understanding of AI and where things are at today, but B, like learn about these cool, like an interesting new developments and new tools. Yeah. Where would you point that person? I mean, I would point somebody to Coursera. They have some great short courses on that just with the fundamentals of AI. And so you can understand AI and just kind of basically what it is, what it enables, what's possible. And then like try and find information specific to your industry. And there's a lot of books there and even accounting, you know, there's books on AI and accounting on Amazon. You know, that's a really good place to start. And for an employee, I think it's ask lots of questions about the existing process. You know, how are we doing things right now? And how can we improve? And, you know, knowing what you know about AI and what is possible, where can some of those, you know, technical solutions fit in to improve things? Yeah. Okay. Interesting. Because, yeah, I think a ton of like a ton of people are very frustrated with the processes at their firms. Like I have a background at one of the big four. And even at that, like I could not believe what some of these processes entailed. So I think like no matter where you are, there's probably great opportunity for this. I think the smaller the place, the more likely you are to be able to actually see this project through and your recommendation actually be adopted. But even at some of these very, you know, the biggest firms, like there's a lot of room for process improvement. And when you multiply that by thousands or tens of thousands of employees, you know, enabling people to do something a little bit more effectively, have a better process in place, I think is hugely valuable. So I really do think that more people need to get that foundational knowledge so that they can kind of just be going through their day to day, looking for opportunities to kind of plug this in and improve processes. So. Yeah, exactly. Exactly. Yeah. So what are like timelines usually like on an implementation like this? Like are these usually very long, drawn out processes? Or if someone wants like a new client onboarding flow, what should they expect? Well, first of all, in terms of client onboarding, they have to do a little bit of work. And that work is mapping out their processes, delivering a flow chart of how their business operates, you know, how the process integrates into every other system as well. So everybody's kind of different, but that's kind of the basics. And then, you know, implementation can take, you know, four to six weeks. It doesn't have to be a huge project. And typically how we like to work is we'll look at the big picture initially, go in and take a look at the big picture just to understand everything and then come up with suggestions. But they're kind of smaller implementations. So it's kind of small project base. And that way, you know, it's not going in and changing everything overnight, right? Yeah. Okay. We're going to go and change this one thing now. That's great. Everybody's happy with it. Move on to the next and be able to kind of measure that ROI each time as well. Yeah. And I think accounting firms in their own way are potentially very well suited to that as well, because there's such a like, you're right. There's a busy season where a lot of firms just kind of shut down and they say, okay, we're not looking at process at all. We're just in execution mode. But then in between that, you actually do have this very long period where, yes, like people should kind of rest, recharge, catch up and all that. But I think if you are a firm owner operator, I think you should really be using that time to kind of consider what can be done better next time. And can I fix that process in this time gap that I've got before things get really, really busy again? I think that's like a really good way for firms to scale while remaining lean. And that's something that we talk about in like exit exit multiples for accounting firms and whatnot is like, can you operate in a lean manner? And then also, are you tech enabled? Are you kind of ahead of the curve of your industry? So I think this is something that pays off in a huge way if you keep up kind of that energy, that productivity, but instead pour it into improving your processes to make the next busy season better. I think cyclical work like that is a great way that you can kind of get the time and energy to focus on these implementations. So, interesting. Yeah, totally. And, you know, selling software services, typically in the past, summer is the quiet time. It's always tough to sell development services in the summer, but it's not that way with accountants. Everybody comes out, you know, May, June, and they're coming up for air. And that's when they're happy to look at things like this. So yeah, it's interesting. Do you encounter a lot of people who are very like resistant to the idea of AI? Or when you're having conversations with firm owners, are they like, very eager to kind of adopt this? The conversations I've had have all been very positive, I would say, you know, the people that don't see the advantages, don't reach out, they don't respond to my messages, basically. But the people that do reach out, there's definitely you can see some similarities with the firms that do respond positively. They're kind of forward looking. And, you know, they're pretty active, their websites are good with a lot of fresh content. So you can tell that, you know, that's important to them, and they're trying to stay ahead of the curve. Interesting. Are there any firms that you see like outside looking in that you think are doing a tremendous job of this that you'd be able to point out? Like you hear about like, big four firm A or B putting in a billion dollars or $2 billion. And there's all these flashy headlines. But I personally, like, I don't know what any of that means. And I have no idea if they're like actually being effective with this. Yeah, I mean, I've, I don't, I haven't worked with any of the big four firms. Yeah, the firms I work with are generally, I would say the sweet spot is like 10 to 50 employees that you can really kind of go in and make a difference. Yeah, yeah, that kind of range. And I think that's like a size of firm where they can make a lot of like bold decisions. You're, you're maybe not quite as deep into like the accounting culture stuff as like the accounting culture stuff as me. But there's, there's this like Alan Colton article about basically how the partnership model is flawed. I think it was Alan Colton. How the partnership model is, is essentially flawed. And that like at a certain scale, just firms become paralyzed because there's so many partners and like nobody can make a decision. And so you end up with like a lack of like bold decision making in a given direction. Whereas like, yeah, in those 10 to 50 person firms, like typically there's probably a firm owner and they'll say, yeah, we're doing this thing. And we're going to be ahead of the curve on this. So, yeah. Yeah, and I'm typically reaching out to partners. I would say 99% of my outbounds are to partners. And I mean, that is also an opportunity for a partner to look good in the firm, right? When they can come back to the company and say, hey, I found this company, there's a bunch of, bunch of different things we can do here to really kind of drive efficiency. Kind of makes them look good as well. Well, that's the thing is like when you're the one to champion this, like very ROI positive project, like I think that's like a, such a good look, right? Internally where I work as well, like we have an individual who really took it upon herself to be like, I'm going to be like the leader of AI in this space. And I'm going to read everything there is to read about it. And she was not in a technical role. And, you know, I think that she really pushed people to kind of look towards this and, and adopt a lot of things. And again, I think that's like a, a tremendous look for a person internally, you know, from a promotion perspective, from a just general goodwill in the company perspective. Yeah. And I think that's a key point as well. If you want to be the champion of AI in your firm, you don't have to be technical. You don't have to even understand AI. You just need to understand what it can do for you, right? I really like that. That's going to be a, that's going to be a clip to share to people. Cause I think that, I think that it is so intimidating, right? But again, you just need to know enough to be dangerous. Exactly. That's, that's awesome. And so kind of the last thing I wanted to touch on is you were actually a big part of building an accounting software with now some AI kind of integration that's gained quite a bit of traction in the market. I'm curious, like, can you talk through some of that? Like what's the mentality when you're building software for the accounting field? Like how do you figure out what they need? And like where there is opportunity to kind of like introduce something better? Yeah. So, so Taxpay came about with three other, three other partners. Two of them were accountants. So the accountant saw the pain, had the vision for kind of a product that could solve that and brought myself in, somebody else to help like build the product. So we would, we would basically, the whole product strategy was driven by the accountants. They knew exactly, you know, what people would do, how they would do it. What, what made sense, what didn't. So I definitely do not know enough about accounting to, you know, go and develop a whole tax payment system. So yeah, partners there that, that was critical. So I would say the team we developed there was, it was a, it was a good fit. We kind of complemented each other. in terms of abilities. Okay, awesome. Well, I've been following Taxpay very closely. I look forward to kind of seeing where that ends up. I think it's a really good product, answers a lot of problems for folks. So congrats on being part of that. That's awesome. Yeah, thanks. It's, you know, at the heart of it, I'm basically a builder. I like to build products, you know, things that help businesses, people. And the funny thing is when I was pitched Taxpay, I didn't totally get it. You know, I just didn't see the value, but I thought there must be something here because these are really good accountants. There must be something. But when I got into it, I really understood. And I remember being on the first sales call and the bookkeeper we were talking to said, I'm gonna give this to all my clients and I will pay for it. And I was kind of taken aback by that. I said, and I said, just out of interest, why would you pay for it? She said, the amount of time this is gonna save me, you know, all the calls to CRA allocating, you know, misallocated payments, just the chasing she had to do. She said, this is gonna make her life so much easier. And that was kind of the light bulb moment for me. Interesting. Is it, it's still Canada only, right? Right now, yeah. Yeah, okay. Awesome. Well, thank you so much for joining me, Julian. I really appreciate you taking the time and educating us a little bit on AI and some actual like practical applications of it. Because again, there's a lot of discussions out there where, you know, one part of the community is saying like, oh my God, I'm so scared, it's gonna replace all of us. And then the other half of the community is just laughing at them and saying, has it taken my job yet? Can it hurry up? And it's really not this binary thing, right? It's just, we can use it as one of many tools that already exist to kind of patch over areas of the process that need to be a lot smoother. We can use it to, again, there's like some new offerings and whatnot that can come from it, but it is really a tool to be used to kind of improve what we're doing. It's not something to be too afraid of yet. So yeah, I really appreciate you taking the time. I'll make sure that I link all of your contact information as well as Zentia's website in the podcast notes. And yeah, I think if anyone is curious about what can be done with AI and their firm, I would really urge you to give Julian a shout. Thanks so much for having me. Peter Pootis