Xero Tax Errors, OpenAI Roll-Ups and Active Accounts Production | Be In The Loop

[00:00:00] Welcome to another episode of the Digi-Tools In Accrual World Podcast, brought to you by The Loop. I'm delighted to say that I'm joined by Alastair Barlow and Eriana, and we're going to be having a great conversation today. Lots of topics. We're a bit zero heavy because we're coming off the back of XeroCon Denver, where there's a whole bunch of announcements and bits and pieces to go through. There's also some really good chat about what's the future of your firms, particularly in this AI world, in terms of how do we blend these two pieces together, but also with that actual

[00:00:30] layer of the acquisitions and PE stuff going on, and some other interesting chats about what we've been seeing recently with the fallout from Sodium and TaxCalc as well, which has been certainly entertaining to sit on the sidelines and watch the back and forth on LinkedIn and various other platforms. So we've got all of that. So Eriana, it's great to have you on. How are things? What's going on in your world at the moment?

[00:00:50] Thank you so much. I'm really happy to be here. A lot, to be honest. I've actually been getting involved in some court cases, financial crime, which has been really interesting. I had them the last two weeks and we won. So it took a while.

[00:01:07] And just in terms of the stories that we're going to be talking about today, some really exciting things about what does the future of our work in the accounting field look like with things like month and close or continual updating of our work. Also looking at what do we actually want from our software? Are we focused on AI or should we be focused on making sure that our tech actually works?

[00:01:37] And also looking at as software is being built, are we looking to take on tech, which is ensuring gaps that we have are being resolved? Or should the current tech we have be able to do a better job with the functionality that technically they could have within their tech? So really interesting conversation. Looking forward to it. Amazing. Amazing. And as always, making Alastair and I look like,

[00:02:06] you know, absolute plebs as well with that massive, incredible setup that you've got there. And us two just in our box rooms, like it's in the corner of the house somewhere. But go on, Al, you can defend yourself now. Yeah, what have you been up to? Well, first of all, great to be back on the pod. So looking forward to chatting with some of the very interesting stories that we've been seeing over the last week or so. Yeah, I've been chatting to a lot of people in the profession and the wider industry,

[00:02:32] understanding and hearing what people are up to, what they're building, what they're tinkering with. So that's been really eye opening. And I think it's if I kind of pull all of that together, it is with the customer at heart in pretty much all that everyone is focusing on and building, which is really nice to hear, not just about efficiency for the firm. It's about putting the climb first, which is really great. So yeah.

[00:02:59] And then aside from that, I've put the bike away for the last few weeks and I've just focused on running. And so I've got a new training block at the moment, which is very much focused on running and getting a new 5K PB. So maybe on the next couple of pods, I'll update you as to whether I've been my target or not. Now I want to know what that PB is. My PB is 1948. Jesus Christ, that's fast. Okay, fine.

[00:03:30] Good stiff. We should go for a run. We should go for a run, mate. Yeah. Well, I'm more like a 25 minute 5K. So there you go. We can do that. We can do that. It'll be good. It'll be fun. Wow. Hey guys, have you heard about Free Agent, the MTD solution that will keep your practice ahead? Prepare your practice and clients to get MTD done with Free Agent's HMRC recognized MTD solution for sole trader, landlord and CIS clients.

[00:03:57] Submit quarterly updates and file final declarations directly to HMRC. Simple digital record keeping with AI powered technology, bulk actions and auto submissions that speed up your admin. Manage MTD compliance for landlords at scale with clean ownership splits and property level reporting. And then not just for freelancers and micro businesses.

[00:04:19] Free Agent scales seamlessly to spotlight what matters and streamline admin for all your clients from startups to well-oiled machines, whether they have 20 employees or two. Perhaps you've got them on our latest episodes of the digital in a cruel world. And Free Agent are proud sponsors of the digital disruptors. So give them a look up when you get a chance.

[00:04:41] So today's first story is about OpenAI's CFO, Sarah Fryer, who has revealed some ambitious goals for its finance function. And that's a zero day close and continuously updated forecasting. The idea is that rather than finance teams spending five or 10 days after the month and reconciling everything, transactions are recorded and reconciled continuously with AI identifying exceptions as they happen.

[00:05:06] What's interesting, though, is that when Fryer joined OpenAI, even with access to some of the world's most advanced AI, she still had to build the finance function and its processes from the ground up. But I think something else that's really important to highlight is that this just shows how AI doesn't fix a broken finance function, right? You still need the good data, the good processes, the connected systems, controls and people who actually understand what they're trying to achieve.

[00:05:34] Only then I think AI can really accelerate that. But as accountants hear this, I don't think they should hear zero day close and think this means removing the accountant in the future. I think it goes further into the idea that what accountants are doing now will be changing. And if AI can reconcile those transactions continuously and flag the exceptions, I don't necessarily want a highly trained accountant spending days ticking and matching transactions, right? Which is what they do now.

[00:06:03] I want them asking, why has this happened? What does it mean? And what should the business do about it? The accountants still own the judgment, the validation and the sign off. So it sounds like the future won't necessarily be about closing the books anymore, but it'll be about continuous accounting with that human oversight. But for now, I believe that every finance function or team should not be asking how do we use more AI, but what are we still doing manually that can be automated?

[00:06:33] And then start thinking about what that AI layer should be. What do you guys think? So I think, so at Flinders we work with fast growth businesses. So obviously OpenAI is on a completely different magnitude to the companies we worked with, but finance always drags behind a fast growing business. So that's no surprise to hear that. I think just the headline of zero day close, I think is nothing really new.

[00:06:59] If I go back 20 years ago and I have been in the profession for two decades, like believe it or not, right? I was seeing 24 hour to 48 hour closes for large multinationals, like clients of mine. That was way before anyone called it AI native. So I think it's been dressed up a little bit as a headline. For me, it's always been a planning and process discipline more than necessarily a tech breakthrough.

[00:07:29] And I think the important word you said, Arione, you said a computationalist and tedious. So I think I'll come back to that in a second. But I think if smaller firms are listening to this and going, oh, wow, that's like unrealistic for a portfolio of hours to close the books within 24, 48 hours or indeed zero day close. I think the constraint there isn't necessarily the technology, it's the review.

[00:07:56] It's like, you know, you've got 30 days or 20 working days of people in a business looking after a portfolio. If you're condensing all that into like working day one to have the final review, that's quite a lot to be able to do with a portfolio versus like in a finance function. But I think the discipline has always been how can we bring as much work within the month? So let's say working day minus five, working day minus 10, whatever. How can you bring that work forward rather than waiting to after the deadline?

[00:08:26] That's always been the case. But I think I am hearing firms, they're able to shrink from, let's say, working day 10 with the portfolio to working day five. And so using Gen.ai and using technology, and it's almost helping firms reimagine what's possible rather than the bias of what we've always done. And kind of deliver a better, more timely service because of that.

[00:08:56] But not get to work day zero. I think as a point of view, I think it's a headline grabber. There's nothing new. But I do think what you said about continuous close, that will be something to shout about. Because that isn't like whoever defined, I can't remember. There is a story about why we are defined as monthly reporting rather than, well, retail dot report slightly differently, obviously.

[00:09:21] But, you know, having like a continuous close and understanding what is my position on the 17th of the month, rather than just the 31st of the month or the 30th of the month. I think that will be the real point, the real turning point for companies to understand the performance at any point in time for the business. And so I think that's kind of my perspectives on that.

[00:09:45] I mean, I'd just add to that, Alistair and Ariana, and say, you know, we've had firms, you know, accounting, bookkeeping firms here in the UK who've been promoting things like daily bookkeeping for a long time, right? You know, prior to that was maybe weekly or fortnightly or maybe even monthly.

[00:10:04] And that all leads into kind of this whole concept, doesn't it, of being able to have exactly what you're talking about, Alistair, being able to go on to your accounting system and see what your performance looks like halfway through the month and understand what's going on. So, yeah, I mean, this zero day close for me, it's like, well, for me, it's almost like too ambiguous to understand exactly what it means. Does it mean zero days after the month end or does it mean zero days when you start that process?

[00:10:32] Because those are two entirely different concepts. You know, zero days after the month end is going to be great if you're a business which has got very simple transactions, doesn't have like long leads on production or shipping and stuff like that. And you're able to track absolutely everything through that business in terms of the inputs and the outputs. If you're a really complex business, you're manufacturing stuff and you're doing other bits and pieces, like actually the accounting for some of that is quite complex.

[00:11:02] And I don't think you're able to do that on day zero after a month end, for example. Having said that, if you're waiting for a couple of weeks because you've got suppliers that are slow and waiting for invoices and you're not quite sure exactly what you want to accrue and so you're playing safe and waiting for that information to come in. Processing that information in zero days is an entirely different thing. And so, you know, the use of this term is like so ambiguous. Like it's just almost meaningless, I think.

[00:11:28] I think there's this thing called a purchase order accrual that can help you out on that. I'm aware of that. But I think the really important difference here between like running a finance function and having like zero day close or working day one close versus running a portfolio is that when you're in the business, you can control.

[00:11:51] Like your client is the other departments, you can control what the client or the departments deliver to you in terms of information. If you're sat there running portfolio, you are that you're one step further away. And therefore, it's that much harder to get the information or the data discipline right to be able to do a working day, a zero day close. But then does that not highlight, again, how much closer we are going to need to be in every business?

[00:12:21] We won't be able to sit there as like this layer on top that just waits for the information. And we'll tell you we're going to have to be involved in speaking to suppliers and get involved in a way that is a lot less traditional. And I know we've talked about this so many times, but being that business partner almost. I think it just strengthens that idea. Yeah. And I think like the first step to that is like a smart finance function or nothing.

[00:12:48] But I mean, in many respects, like what you're saying there, how are you going to about getting closer to the client, right? And being getting closer to those inputs and outputs. That's not necessarily a technology issue, right? I mean, you know, Alistair will promote his smart finance function until he dies. But like, you know, much of what he was doing with the guys at Flinders and stuff was, you know, it was about having that capacity to be able to engage with a client in the right way and understand exactly what's going on day to day. Yes, of course, technology helps.

[00:13:18] It makes it easy to transfer that information between your parties. It makes it easier for you to be in the business a little bit more when you're sitting in zero, et cetera. But I don't think any huge, large, complex business, particularly if they're running out of ERP, is going to get close to this. You know, we sit as accountants in this world in very different spaces, I think. And I hear what you're saying about your experience at PwC, Alistair.

[00:13:42] But even so, like big businesses have like so much volume and so many things that could potentially go wrong. You know, this zero-day close is just ambitious, I think. I think the bigger, don't forget, the bigger the company, the bigger the materiality. So, arguably, it's much more focused on controls.

[00:14:06] The controls being in place in the business to get it materially correct rather than taking and chasing a bunch of stuff, right? So, should we move on from, unless there's anything else on that, should we move on from like one end of the spectrum as like, I don't even know what OpenAI is valued at these days, to the other end of the spectrum.

[00:14:25] So, huge, huge tech company through to Prosaic, which is a new New Zealand company, tech company, AI native company, which is dealing the opposite end of the spectrum for firms with what looks like a lot smaller requirements from their accountants. So, Prosaic is an AI native ledger from New Zealand.

[00:14:51] I think it's only available in New Zealand, maybe Australia, but definitely not in the UK. So, you look at this, pricing is $10 a month. Now, there's been a lot of pricing chat on social media around four price hikes in the last 12 months or so. Someone can correct me if that's not right on other products focusing on the same market. So, some of the angels are quite interesting here. We've got Michael Woods from Receipt Bank, co-founder of Hubdog, co-founder of MYOB, founder at A2X.

[00:15:21] amongst many others, investing in this and backing this. It's a cash-based general ledger. So, limited kind of like determines what their initial ICP or market looks like. I think it's cash-based. Again, someone can correct me there. So, it's got limited scope, I guess, and very clear where it's clearly defined, where it's focusing.

[00:15:46] But what I think is really interesting, if you watch a demo of here, is kind of reimagining what a general ledger should or could look like, which is what I really like. It's not just kind of making something faster that's already there. So, it's got these chart templates, which I think you can set up a template and then you can update it and push it to all its kind of children that you work with.

[00:16:14] So, for example, when we're working with, let's say, e-commerce businesses, if we wanted to tweak a chart of accounts with one of like a template level and we wanted consistency and standardization across all our e-commerce businesses, we can push that across the entire estate rather than having to do it one by one, which I think is quite a neat little thing there as well. And you can also apply rules across your entire workspace. So, I think whether that is you apply it to a rule to a holding company and then it pushes it down to all its subsidies,

[00:16:44] or again, whether you can push it down to kind of all your companies in your portfolio. Obviously, being a new general ledger, it lacks a bit of an ecosystem in terms of like apps in the marketplace. But what they do lack there, I think they're making up with by having a completely open API and able to integrate or connect up to an NCP.

[00:17:05] So, you don't have to wait for vendors to build the integrations and probably build quite poor integrations for what we actually want as accountants. And so, you can build it yourself, which I think is a pretty cool way of doing things. So, yeah, the whole Bankbrek side of things also looks a lot more slick than what we've seen in others. So, I quite like this.

[00:17:29] I don't really like part of their product where it's got these like, what it looks like, if you've ever seen a vide-coded website, it's got like this sliver of a stroke on the left-hand side of a box with kind of like a slight curve and it's got the same colors. It looks and feels a bit like that. So, I don't know how quickly this turned out some of the design with vide-coding, but that's maybe one part I don't quite like. And I think one interesting thing on the website, which I think maybe we'll come to a bit later in the UK market,

[00:17:57] is that they say they're custodians, not owners of data, and that financial data is sensitive and is not theirs, and firms and clients stay in control of their information. And so, yeah, let's revisit that in a bit. But I think this is an air of thought and challenge about it, which is, I think, what we're missing in steady state general ledgers at the moment that aren't really coming up with anything new. So, yeah, I don't know if you guys have seen this.

[00:18:27] I think we obviously were, I think if you're in the same, I know you are, John, Arion, and I think you might also be in the same WhatsApp group that had a little bounce and back this morning. I think the more I think about it, the more I think, the more I question, are Zero just going to be a legacy? So, I don't know. What do you guys think? Have you seen this? You played by it? Before I really jumps in, I'll maybe say, I mean, you know, there are some interesting people behind this,

[00:18:55] but Nick Holdworth in particular is a sort of a former Zero, I think, GTM, if I remember rightly, also worked at Vend and Tradify, so kind of familiar with the ecosystem broadly as well that Zero sat in. So, you know, he's definitely got some chops about it in terms of like, you know, knowing the space and everything else, which is interesting. I also looked at the website, and having Vibe coding some stuff myself, I saw that same sliver thing that you saw. It was like, oh, shivers.

[00:19:24] That kind of crops up in everything that I built to begin with, and so you kind of like come up your own design system, which is hilarious. I mean, I think I haven't really dug into what Prozac is trying to do from a GL point of view. I haven't really looked at the screenshots and that kind of stuff properly. But it's such an interesting thing, isn't it? Because, you know, I think all of us have probably been approached by someone who says, oh, I've got a new idea for a GL in the marketplace. And then, you know, personally, I really struggle to sort of say, what is it that I want from a GL?

[00:19:54] You know, whether it's Zero or QuickBooks or something more complex, that isn't already in the market as far as I can kind of go and get it in the ecosystem. And in many respects, for me, that then becomes a bit of a challenge, because at the same time, I'm saying I don't want these products to kind of merge into the ecosystem products and kind of like just spread themselves too thinly in terms of capability, because that then means that I'm compromising on the quality

[00:20:21] of like what those products are doing in terms of like workflows and information delivery and the effectiveness of what they're trying to do. So it's a strange thing. But that's kind of like my perspective on this at the moment. I think it's really interesting as we see so many of these kind of softwares now coming up, whether they're niche, tiny little softwares dealing with only one issue or the bigger kind of GLs.

[00:20:48] There's so much choice and options every day. What does that mean for the accountant when they're choosing, when they're trying to run their business? What do they go with? Do they stick with one and consider efficiencies? Or do they look at which GL may be more appropriate for a different kind of client? But either way, I think there's just so much choice now that it's going to become harder

[00:21:16] and harder for accountants to try and keep their costs and efficiencies all kind of in order. That's a really interesting point, because you get two sides of the equation. You've got the client, what's best for the client in the client's eyes, but also probably from the accountant, what's best? Because they put a client first like that. But then also the accountant is running a business and what's best for the accounting firm as a firm

[00:21:45] to run an efficient business. And those two with the, I guess, the proliferation of every other day on LinkedIn, there's a new AI native general ledger that has just raised $100 million on a billion dollar valuation. And so the market is just expanding, whereas previously there were like three, four, five, whatever. And broadly, it would be a sensible choice for who the type of clients you serve would use and you as a business. And there's a nice match there.

[00:22:15] But I think it's going to make it harder and harder. So I was thinking something yesterday that they would have been putting everyone on zero. And now a few of the AI native clients that they are speaking to, they are wanting to go with an alternative provider. So they were looking at Ramp, who've come to the UK. And so that kind of means that they have to be able to two softwares, two solutions. And they're a small business, a very small business, and they're growing.

[00:22:44] So you can see as they expand and as they grow, are they going to need kind of specialism in more and more general ledgers or ERPs? The other thing is, though, I think with this, the perception from both businesses, so clients and some accountants, is that now all these softwares are coming up. They've all got that AI layer in it. They're all saying that they can do most of everything themselves, right?

[00:23:10] I had an interview this week for a head of bookkeeping within our firm. And I asked, you know, what kind of software are you using? Have you ever done? When was the last time you did a VAT when you reviewed things? And he goes, oh, I don't need to. I was like, what do you mean? He was like, AI does that every day, doesn't it now? Like, why do we, we don't need to get involved. I was like, what about Bankrec? You don't need to touch a Bankrec. The system does it all.

[00:23:37] So his understanding was that we don't need to touch absolutely anything. I mean, he talks himself out of a job. What did this person think the job was then? Just to sit back? Literally. I was like, so, you know, would you get involved in bookkeeping? He was like, no, I don't prefer to. I'll just do the like year-end journals. I was like, okay, so I think we've got a misunderstanding here. But just the idea of, and that's an accountant who's got 15 years of experience.

[00:24:05] Imagine a business owner who is seeing this messaging with every new kind of software that's coming up and saying, we do all of this for you. Your accountant doesn't need to get as involved now. What is that going to mean for us? How do we say, no, honestly, we are really needed. We still need to do this, although it can help in other ways. But we still need to be involved in this. I think there's that worry there with, as we get more, the messaging is just all over the place. It's really a difficult one, isn't it?

[00:24:34] Because whether it's right or wrong, as in whether it does or it doesn't do it, if you spend enough on marketing, people will believe it. Absolutely. Until it goes wrong. This isn't a new problem. Yeah, this isn't a new problem, though, is it? Because we had this in the early days of Xero and then some of the new apps as it came into the space, saying, we could do everything for you. And probably, I mean, you've probably seen this more with QuickBooks. Like QuickBooks at one point were almost saying, you don't need an accountant or bookkeeper in the UK. And obviously, it annoyed a lot of people.

[00:25:03] And so we've always had this kind of like your battle, if you like, of understanding what's going on between the actual reality of the technology on the ground, what the marketing and the salespeople are telling the world, and then what we're also seeing and what we're also capable of delivering as well. So, yeah, it's nothing new. I think the defense from the marketplace is still the same. You've still got to demonstrate value.

[00:25:30] You're still going to have clients that come to you who completely mess up because they've just assumed that the technology is doing it right. And then you'll charge them a lot of money to fix it. And if you don't charge a lot of money to fix it, you're a mug. And, you know, that unfortunately for some people is a lesson that's learned in an expensive and difficult way. But that's just the way it is. I would say the one challenge that we have as a market, you know, as a marketplace, not just this isn't just in the UK, but it's broadly around the world,

[00:25:56] is that I still think that most of the accounting bodies don't speak with a singular voice and don't actually articulate particularly well the value of accountants and bookkeepers and tax advisors and everyone else of what we're doing. And they're not particularly good at being able to push back against, you know, these huge amounts of money that are spent marketing these technologies just because simply they don't have that capability, but they also don't have the ability to kind of communicate in the same kind of dynamic way as well.

[00:26:26] I will move on to the Xero Global App Awards, which they announced, what, six days ago? What was that? Last week at the tail end of XeroCon in Denver. Oh, dear. So they announced a bunch of global category winners and then some regional ones as well. So I'll whiz through these. I won't go into these in too much detail.

[00:26:53] Maybe on the newsletter and stuff we'll kind of explore some of these in a little bit more detail. But some will be very familiar to you. Some of us are ones that I have not heard of myself, partly because they're not UK-based. So the Global Small Business App of the Year winner was Goji. The Global Practice App of the Year was Ignition, very familiar to us all. Global Innovation App of the Year was ShiftCare. The Global Emerging App of the Year was Pulsify. The Global People's Choice Award is Mayday.

[00:27:21] David Tuck's always good at getting people to vote for him. And the Global Development Partner of the Year was Ahoy Ahoy. And then on the regional basis, Small Business App year winners. Australia was Sin7 Core. Not sure that's a small business app, but anyway, we'll see. UK was Approval Max. USA was A2X. New Zealand was Paid Nice. Canada was A2X. South Africa was Paysoft. Asia was Simple AI. And Ireland was Simple Pay.

[00:27:48] And then the regional practice app of the year winners were Expert in Australia, Socket in the UK. So kudos to Johnny. Sweet Files over in the US. I'm not sure that's a... Well, I guess it was a practice app. It's next to the topic of management. Mayday in New Zealand. Content Snare in Canada, which is interesting. The Portal Junior in South Africa. Tofu in Asia, which I've never heard of. And then Fuse sign in Ireland. So that was the roundup of the winners. You ruffled that one off, didn't you?

[00:28:14] I think, for me, one of the interesting things is that both the global app and the UK app, I can't remember the categories, same categories, were both proposal tools or proposal-ish tools. Which I thought was interesting. There's so many pain points in a firm that proposals, onboarding, is that the biggest challenge in your firm, Mariona? Why do we think proposal tools won? Personally, I don't think proposals are the biggest one. For us, I think it's still chasing.

[00:28:44] As in chasing? The data records of clients? Yeah. Yeah. I don't think that's something that anyone will be able to get right for a long time. Yes, I'll fix it. I've already fixed it there. I know, we have it. I think it depends. Actually, here's just a bit of market research.

[00:29:03] Would you be happy for some sort of application to go into all your clients' emails and fish out all the invoices and documents that you need? Or would you think that's too invasive? No, if it's within the ecosystem. I don't think there's anything wrong with that. Into their emails? Oh, no, no. Not into their emails.

[00:29:26] But if there was something that can look at going to the GL and see what's outstanding, what's missing, which invoices are not attached, all of that. Then go back to the client and say, hey, you still haven't sent us this. I would love that. Because what we're doing right now is we're wasting time and asking for information. Once we've checked what's already come in, we're checking what's missing, sending those to the client. They'll send 10% back. Then we have to look through it again.

[00:29:54] Then we have to send them the same list again. So this happens over and over and over again until we get to getting all the information in. Now, if that could be automated by someone doing that for us. It's the client that's the real problem here. Sorry, I probably didn't do too much market research on your podcast here, John. But I think it's an interesting pain point that that is the tough part.

[00:30:14] Just on some of those categories, I spoke to three people in the last week that are building their own either proposal or onboarding tool because they want to own the experience. And actually, I've got an article on News West coming out this Saturday. I don't know if this pod will go out, but it is about customer experience being like the biggest differentiator in an accounting firm. Because we all deliver a VAT return or whatever those deliverables are. And we charge a price.

[00:30:42] And the thing that makes a difference is on what the price we can charge is customer experience. And so the firms I spoke to are kind of like, we want to, first impressions count. And therefore, we want to own that experience, which I think is quite an interesting thing. I think that's a really interesting take. And yet, isn't it so funny that that's obviously a technology play from those people you've spoken to.

[00:31:08] And yet, we continuously talk about how the human experience and the relationships are the differentiator. So one of those cannot be true whilst the other one is true. So something doesn't quite give that in terms of if we're looking to give good client experience, is that a technology platform? Or is that the people behind the technology platform? Well, I challenge that. And I'd say in an accounting firm, it's both at the right points in time.

[00:31:38] If you load up Netflix, right? And it's got an algorithm that tells you like what you should watch and blah, blah, blah, right? That experience of being able to flick through is actually quite a nice customer experience, isn't it? And that's just pure technology. There's no one to speak to there. So I think it's a bit of both at the right time, but knowing when the right time to have the human interaction. But I think the point there is just being able to own the experiences is what they wanted.

[00:32:07] I'm not saying it's right or wrong. And I think it would be different for a different type of firm. And I think it would be different for a different ICP. But looking at the entire end-to-end spectrum of what the client journey is from onboarding through to deliverables, depending on what your ICP is and what your firm is about, you should pick apart and go,

[00:32:34] well, actually, I'm quite happy for off-the-shelf technology there because it's great, it's good enough. Or I want to inject my own personality, my own brand, my own experience there. Or I want to have a human touchpoint or a human face-to-face point or whatever it is. But I think the first part is most firms don't map that out, right? No. And actually, I think probably the positive thing to take care of that is that those are organizations that are looking at that kind of experience.

[00:32:57] And I still think too few accounting firms think about the customer experience when they're thinking about how they deliver things, how they interact, how they bill, how they do all of the kind of operational mechanisms of an accounting firm. It's all about how can I make it as seamless and efficient as possible for my team? And how can I maximize margin? Maybe not always, but it's normally about how can I make it as efficient as possible. That's what the driver has been in the last 10 years.

[00:33:27] And probably will continue to be with private equity, right? Because they're not, they're probably, maybe they pay lip service to it, but they're probably most of them not going, it's client experience first. They're probably going, how can we fucking jack up the multiple here, right? I think very few of them will be thinking about client experience, right? Because that's not in the modus operandi of how they're trying to achieve their success, right? They are looking for an efficiency play to increase the bottom line and get out.

[00:33:54] Maybe in the next few years, maybe they'll start to switch that if we see the erosion of value as a consequence of technology impacting. One thing I will just say, just to go back to Mariana's point about that challenge about going back and forth with the clients for information and data, is I agree with you as well, and it is a challenge. And yet, ironically, in the audit space, this isn't solved, but it is better than it probably is in the bookkeeping and broad accounting space,

[00:34:20] because this kind of like, what we call PPC, pervert by client information stuff that you get in audit, is an area that's had a huge amount of focus, probably at the upper ends of the market in terms of firm size. And there are some really interesting pieces of technology there where you will send a request list out to a client through some kind of platform or portal.

[00:34:45] Well, you know, the client can then respond to that upload information and is using things like AI to validate whether that information is what's been asked for. And if it hasn't been asked for or hasn't been delivered properly, then it will go back and it will start to chase some of that information and prompt them to provide the right information. And because it's quite quick, because the AI has been able to extract information from the documents or the information that's been uploaded, it's quite effective because you're not losing the client at the point of kind of like sending that off and then having to wait for it to be reviewed to get a response back.

[00:35:15] It's not quite immediate, but it's kind of, it's within a minute or so of people uploading stuff. So a client can see like, hey, like there was 25 requests, 20 of them are green and five of them are red, either that's because they're late or they've not given quite the right information or it's incomplete or something. There still is a human check in a lot of these things, because obviously if you're asking for a lease and you're not specific and someone just uploads any old lease, it could be the wrong one. So it's not, it's definitely not perfect.

[00:35:44] But it is interesting that you're like I say, this seems to be getting closer to being fixed in the audit space, but it has that technology has not dropped down into all the regions of all their areas of accounting practice. There, again, I wonder if someone could create something really interesting, which is sort of like your bot light that sits in your emails and says, oh, Ariana, look, there's an invoice there and you need to send it off to your accountant. Maybe you want to sort of like get this stacked up and ready to go.

[00:36:13] Without it being too invasive, there's a technology that you can deploy out to someone, you know, kind of keeps the walls around the privacy of your inbox or whatever it is that you're using, aside, but that helps to kind of like, you know, fill, fill that gap between those two organizations. I feel like that's already kind of here, though. You can do that with things like Claude, for example. You get it to sit in your inbox and it will identify those things for you.

[00:36:42] And in some cases, it can actually forward those things on to the accountant. But yeah, it's not, there's not like an out of the box ready solution. That's the problem, isn't it? Because like for you to deploy that with 50, 100, 200 clients or something, you don't almost have to sit down with them every single time and say, hey, have you got a Claude subscription? Okay, now let's walk you through getting a Claude subscription. Let's walk you through setting this up. And you could probably do it with a mixture of skills and bits and pieces like that, but it's not seamless, is it?

[00:37:10] That's the thing, you know, like I said, it's not out of the box. So that's the challenge. John, who are you seeing doing it the best in the audit space? In the audit space at the moment, so ironically, I mean, Caseware did a fairly recent acquisition. I think we covered it on the pod a while back and they bought a product, which have now started to build into their main platform. And that was very good. Datasniper also bought another platform as well, which they've started to roll into their broader products.

[00:37:38] Those are, for me, probably at the moment, they're the two like leading, you know, like PPC technologies as we would describe them. And there's a whole bunch of others in the marketplace who are now behind them as a consequence, who are kind of like reinvesting into their platforms, your inflow circuit, a bunch of others in the US. Well, talking about fixing things, I'm going to switch things up a little bit and move on to Xero. So understandably, after XeroCon, there was a lot of excitement with, you know, what they're planning to do and what the software can do.

[00:38:06] But one accountant decided to focus on something much more kind of basic, and that's proof checking a set of accounts produced through XeroTax. They reported finding several issues, including things like incorrect directors on the cover page, an inappropriate section 444 disclosure, duplicated wording around finance leases and higher purchase, and limitations around disclosing prior year restatements. These raise the important question,

[00:38:33] are we becoming so focused on the AI race that we're forgetting about the fundamentals? And I'm a huge advocate of AI, right, and accounting technology. But before accounting software gives me this amazing, shiny AI assistant, I actually needed to get the accounts right. If it saves me, let's say, 20 minutes, but then I have to spend that time checking whether the accounts and disclosures are actually correct, then what have we actually achieved here? But I think the responsibility probably lies on both sides.

[00:39:02] Software providers do need to keep innovating, but AI can't come at the expense of core accounting functionality. But at the same time, accountants can't assume that because the software or AI produce something, it must be right. We're still responsible for reviewing and signing off the work. And that trust with the AI will be, once we've got the boring things correct and consistently.

[00:39:28] Ultimately, I think putting brilliant AI on top of unreliable accounting infrastructure doesn't solve the problem, but it just allows us to make those mistakes faster. And it's something that I'm seeing all over LinkedIn in different user groups, pretty much every comment that you'll see with every announcement that vendors put out to say, we've got this new feature or that new feature. The comment is almost always from a number of accountants saying, thanks, great for this,

[00:39:55] but can we actually have another feature that is there that's not working, being fixed before we move on to something else? So I thought that was super interesting to highlight. We do need to focus on the basics because what's the point, right? There's no point in having shiny software if it's not doing things correctly. I would agree with you on that. I think there is a responsibility on my side,

[00:40:22] but this is one of the reasons why the likes of CCH, Caseware, Digita, Sage, Iris to an extent, still dominate the accounts production layer in the broad accounting market because they've got the domain knowledge, right? And anyone else competing with them, whether it's Xero, whether it's Silverfin, whether it's a bunch of others that are in the marketplace, do not have that domain knowledge. And so they are not capable of producing a piece of software that does the things that we want them to do accurately,

[00:40:52] reliably, et cetera, because they just don't know how to do it. Fundamentally, they don't understand the technical requirements of getting a set of accounts out. It's interesting because I think this piece came from Andrew Coulson, who I know reasonably well. He's an ICAW member, been around for a million years, like I have, and he is quite a technical accountant. So it's not too surprising that he's going to have identified and pulled up some of these things. Interestingly enough,

[00:41:17] I was doing some accounts for one of the businesses I run on Xero Tax the other day, and he kept duplicating the bloody corporation tax amount. And I was like, what the hell is going on? Like, why can I not get this right? Because every time I run through the accounts process, I'm just getting the P&L and the balance sheet, just don't agree. And what I'd realized was that because when you do the combined filing with Xero Tax, because it's calculating the corporation tax for you, I'd already calculated it and posted it in the trial balance.

[00:41:46] And then it was looking at the corporation tax comp and then adding the two together. And I was like, this is a page. And in the end, I figured I had to fix it. I started to move. We had some like under and over tuition from the prior years. I had to move that to another TV account. Then I had to redo the XBRL tagging, and then it got it right. But it was like, from a user experience point of view, yes, I, and that was a problem that I should have known about and should have understood properly, but it wasn't a good experience for someone who's got experience to go through and try and fix that. Because yes, I've been partially lazy in the way

[00:42:15] that I've set up my chart of accounts and what I've posted to the TV, but it should have been really easy to kind of like, just fix that problem because I had calculated the tax and put it in the trial balance. I didn't need it adding on twice by the software trying to be clever. So, yeah. So, I mean, look, as you say, Arionia, the responsibility is always on the accountant right here. Like, we are the technical experts. We should be able to get these things wrong. And it's all well and good that we go out and blame technology, but I've seen plenty of accountants produce shitty accounts

[00:42:45] which are not reliable, not accurate, and have got things wrong in them. And you can't blame the technology for that. You can only blame them. And yet, we never really call them out, do we? I mean, I wish we should and could, but we very often don't do it. I spoke to an accountant recently. I was reviewing their work from another practice. And I was like, these figures don't make sense. He was like, but they're nearly accurate. The bank nearly balances. So, it's fine.

[00:43:17] Should we stick with zero a second? And actually, Arionia, first of all, I think when you did your seamless shift before from John and I to like your perfect anchoring and I thought that was like uber professional. So, well done. Well done for that. When you talk about zero, you're obviously a QBO or a QuickBooks only or predominantly business. How does it feel when you talk about zero?

[00:43:46] Does it make you want to trial zero again or not? Listen, we use it for some of our clients anyway. So, where they were large businesses, going through and moving them, it's just too much of a job. So, we do have some zero clients. If I'm honest, right, and I know everyone's in like all kind of camps, but they're all pretty much a similar thing with buttons in different places. But we made, you know,

[00:44:14] we just made the strategic decision at the time. We'll stick with one. And it was the one that was giving us the most support in getting set up. Whereas Zero just sent us a link and said, here you go. QuickBooks actually held our hands throughout the whole process. And that's something that I think they still have even to this day. Zero is just an email and they'll get back to you in 48 hours. So, that was, you know. Otherwise, yeah, they're pretty similar. Well done for Kibia, for Intuit,

[00:44:44] for building that great customer experience with the firm as opposed to Zero's, which was the pure tech part, which obviously didn't work in that such point. But if we do stick with Zero, so Zero had some bank break updates. So, I don't know if you guys have used the auto bank break, sorry, the auto bank break functionality in Zero. A few people have been quite critical of this in the past on social media. I'd never used it before, the updates,

[00:45:14] myself actually. But I have used it since the updates. I think it's only available on growth plans and above. But some of the changes include a reconcile page to making reviewing easier. So, on that page, you can see whether you've manually reconciled something or whether Jax or whatever underlying name you want to give it is making the decisions and kind of what the decision they've done,

[00:45:43] whether it's reconciled based on a match, whether it's reconciled based on a rule, a memory or a prediction, or even manually. And you can go into that page. You can make corrections from that page unless it's for a bill payment. But just a side note, just a side note on bill payments. When did we become so dumb as accountants to call invoices bills? I just, I just, we've all got into this realm of calling them bills. But anyway,

[00:46:13] so it acts when it's confident, right? So, so that on the face of it sounds good. So I like this page. I like being able to see it. I like the fact that we can go in at a glance and edit it. But what, what's confidence? As somebody reviewing something, we want to know, I personally, I'd be interested in how you guys look at this. I want to know, well, what's, what, what confidence level are we at? And what's,

[00:46:40] what are the inputs into getting that confidence level? And that's what worries me, not really knowing what that black box is doing. So if I, if I look on one of those memory, for example, I can, I can do a little hover over it and it says, Hey, in the past you've reconciled similar statement lines in this way. Cool. What is the past? Is it over the last week, the last six months? How many have you seen? How many have been different to this? Have you seen 80% this way,

[00:47:09] 20% a different way? Have you seen it? And so that, that for me, like makes me a little bit nervous as, as a controlling accountant, right? That wants to know why a decision is being made. So I think good changes, but like, I want to see a methodology. I want to see a standard across the industry. As to what confidence is, how memories are applied. Um, and we're not getting that. We're getting tech companies interpretation of, Hey, we're confident. We'll do it this way.

[00:47:39] Yeah. That, that that's good on the face of it. What, what does it really mean? Um, and I think as we're going from like this deterministic world to probabilistic world, I think how that's been calculated or, uh, built up is quite important, certainly to me. Um, but yeah, you guys like this, uh, or not. I personally like it. We're using it in, um, also receipt capture, not just like bank rec, similar kind of methodology.

[00:48:08] I think it's amazing. And I know there is that risk as you're saying, like, how did it make that decision? Where is it coming up with that? Okay. We have to, it has to, uh, one software I'm using, it has to see, um, the information at least three times for then it to say, right from now on, I'm going to put it here. But if that's saving up time from categorizing everything myself to then moving and shifting to, okay, now I'm going to focus on accuracy. So I'm reviewing things. And the more I review,

[00:48:38] the more confidence I guess I have in the software, maybe that's dangerous because the software one day could decide to change something. And if we've already thought in our head, yeah, there's always had it correct. If it suddenly changes that, I may not notice it. But when we're finalizing accounts, we should be able to kind of notice like we would review any team members work that, you know, has come for us to review. It's the same, similar kind of thing.

[00:49:06] But I just think it just shifts what they're doing into learning about or focusing more about accuracy and then starting to hopefully think a little bit more about why the business has done things a certain way and start to think critically about how they're running their business rather than keep categorizing everything manually. Yeah. I mean, I, I, I think I fall somewhere in between the two of you. I mean, every owner, I think you're absolutely right. It's like, I've,

[00:49:36] I've used the auto bank rate feature, you know, and, and it's, it's fine. Um, I don't have it auto bank, also reconciling everything just because it's either just not getting things right all the time, or, um, actually sometimes it's just not actually doing any of that auto rec for whatever reason. And probably cause I haven't said to as well. So there's definitely a user element there. Um, what I'm interested in though, is like this kind of like this concept of, uh, you know, like confidence, right? You know, when you, and I think you kind of like, you,

[00:50:06] you, you nailed it on the head area when you're saying like, you know, I, I use it in the same way that, that, that I use the output from a human, you know, in my business as well is like, I'm reviewing this and looking at whether it's reliable. And yet when you, one of your human staff is posting stuff in zero quick books or whatever, you don't ask them to put a confidence score next to it. Do you? You don't ask them to say, Oh, do you, you know, do you, was that a hundred percent accurate? Or were you just kind of like taking them out of a chance and hoping I wouldn't spot it? Like, you know, we don't do that. And,

[00:50:34] and what I find really interesting, you know, when Alessa's talking about this, it's like, you know, why, why are we so like absolutely in the weeds with confidence scores around a new piece of technology, which, you know, we know it has certain limitations, but those limitations are almost human in many aspects. And, and yet we're not applying the same set of rules, you know, like we're, we're asking for confidence scores, which I think Al, you're almost saying is like, in some respects could be a bit meaningless anyway, because like being a hundred percent,

[00:51:03] you know, I'm a hundred percent confident about how successful, how good a skier I am. Right. That doesn't mean that I'm at the Olympics winning gold medals. And, and so, you know, what's confidence being, you know, in terms of like the output that we're expecting from people, you know, I, I could have told you when I was a trainee, I was a hundred percent confident with the output that I was producing every day. and yet the managers and the partners were tearing it down. And so, I mean, I was doing a crappy job and I was having to learn again, how to, how to do it better. Yeah. So I think, I think we're, I think to some extent we're in violent agreement, but,

[00:51:32] but I think maybe you've misinterpreted what I was saying. And so that, that's either on, on the way I've explained. That's your, that's your fault. Yeah. That's definitely your fault. So, so they're the ones that give confidence score, but I was talking about inputs. Like what's the methodology to be able to post something somewhere. Right. And so is it, we've seen it once, therefore we'll post it there. We've seen it twice. We'll post there. We've seen it eight times to this cat screen, twice to here. And you're right. A junior would be like, I'm a hundred percent confident, but you know,

[00:52:01] maybe they would say that maybe they go, Oh, this one, I'm not sure, but they'd be confident to do it. So you're, you're right to kind of challenge the concept of, confidence. But I think the difference with a trainee is you will, you will coach them into, right. If we see it, if we've, if, if, if you've seen it here, the last three times, five times, six times, So you're the one that's giving guidance to how to treat a certain cost or coding based on the patterns that have been seen in the past.

[00:52:30] We'd have no visibility on that in zero in the bank, To understand, is it, are they seeing it three times? Are they seeing three times in a certain amount, a certain period? And what happens if there's a fourth time that somewhere else, like if, if, if a junior sees it. Isn't that just AI shortcut in the process though? Because like, I mean, we, we talked about this before, because this used to be a problem with, with decks or what was receipt bank, right? Is you, you, you buy, you buy toilet rolls from Amazon every week for your office.

[00:53:00] And then one day you buy a printer. And the problem is, is receipt bank still sticks it in like office sundries or something. And it's like, no, no, it's a capital item. This isn't, this isn't a new problem. Like, you know, in terms of like, how do we, how do we ensure that your transactions are categorized in the right place? And consistency of posting is only one of those elements as an indicator. The other thing that I, I've seen a lot of in terms of complaints about, you know, the auto bank rack feature is that, you know,

[00:53:28] if you have lots of regular invoices from a, from a supplier or a customer, and then those bank transactions coming in, you know, the AI is reconciling against the most recent invoice, not the historical invoice for the, for the same amount, which is kind of the, the normal logic you would kind of assume. Now, so, so these kinds of things are all just things that as a trainee and as a bookkeeper and as an accountant, you learn because that's the kind of the way that the logic applies. Or sometimes if there isn't logic,

[00:53:56] you just start to apply your own in terms of, well, that customer is probably not paying the most recent invoice. They're probably paying the historical invoice pitch in, in an absence of information to clarify that. And so this is just part of what we need to train into the models, right? Isn't it? Surely in the same way that we train us now. Yeah. So I think we're in violent agreement. So that's exactly what I was saying. So you, you give guidance and input into a team member. So they execute based on your guidance. We don't know what the inputs are into this model. Because it's not our model.

[00:54:28] I've got no idea what they're matching on. Yeah. But I guess, is this not a thing, sorry, is this not a thing why these things will start to learn as they get more context from either that particular organization or for that particular firm? That's just an awful question. Hopefully, but we don't know. We don't know what the ingredients of those are. So ideally we need to understand what is the workload that is following in order to identify what's going on,

[00:54:57] which is very interesting because WorkHero have just launched their work functionality for professional firms, including accountants. And actually there's a lot here that I actually really, really like. The idea is that you build a process once and the system then manages that workflow consistency, consistently across your client base. So for example, with year end accounts, work can move automatically from preparer to reviewer back for amendments if necessary.

[00:55:25] And ultimately to the client for approval with documents, communication and an audit trail, all staying connected. And they've also introduced an AI obviously generated handover summary. So when a job moves between people, the next person can quickly understand what's happened without having a peak to kind of piece everything together. From an efficiency perspective, I think this is like really amazing development. However, my question is,

[00:55:53] do we really need another piece of software to do this? I think WorkHero has designed something to, well, it is designed to sit alongside your existing practice management system, but surely a good practice management system should already be managing the workflow for the practice anyway. It should know who's responsible for a job, what stage it's at, what's outstanding, whether we're waiting on the client or who needs to review it and what needs to happen next. That said,

[00:56:22] I don't see it as a criticism of WorkHero. They're actually solving a genuine problem because many practice management systems simply don't do this well enough, if at all. And these are, I guess, exactly the areas where accounting practices lose huge amount of time, which we've discussed earlier today as well. It's not necessarily doing the accounts, but in the handovers, approvals, chasing emails and figuring out where a job actually is. But for me,

[00:56:48] I think the future should be less about adding another app where we find a gap. I would love one intelligent practice management ecosystem where the job, the workload, client communication, documents and approvals are all connected. And then if they're all connected, then you can have that AI layer sitting across all of those to identify bottlenecks and to actually help manage the practice proactively. But yes, I like what they've built,

[00:57:15] but I think it just highlights that practice management systems need to do so much more. So, I think work papers have launched something that they have probably been talking about for a really long time. I know that they've been testing it with a number of firms here in the UK for, for gosh, months, maybe even a full year. I think they've been going quite deep on making sure that this product's really sound. So they've, they've launched account production into their products. So this is now an extension of the work papers solution that they have.

[00:57:45] What they've actually launched is FS102, which is, which is really interesting because I think it's one of the few products out there in terms of the new challenges that does full 102. They've also released 102.1a, FS105, LLP, sole trader accounts, no mention of partnerships, or any do this sort of like weird and wonderful stuff that people request. But that's fine. Cause I actually think focusing on like the core, like your financials, cause that's always the interesting stuff from, I'm just like, it's edge cases,

[00:58:15] which can come later. But effectively now, I guess what this allows active to do is probably like compete really directly with Silverfin, effectively. Which I think has been a sort of a long-term ambition for them. They've certainly targeted the upper end of the marketplace in terms of the size and scale of firms. And, and so that's what they've got. So you can effectively pull a trial balance from a bookkeeping system. They've got lots of really good connections and stuff. Then obviously build that out in the work papers tool.

[00:58:44] And that then flows through out into the, into the accounts production side of things and all the usual stuff that you would expect in terms of, you know, tagging, filing a company's house, bits and pieces. And you can, you can't file directly from the software. So you're having to go through a third party. So whether that's like the company's house, web filing or using a product like Informal Direct, that's kind of the only limitation, which is an interesting one because having, having built stuff using the company's house APIs,

[00:59:11] they're very accessible and actually relatively easy to utilize in terms of like getting some of that other functionality in there. So I would expect that that might come down the track fairly soon. And, you know, this, like I said, just brings them up to parity with a lot of, a lot of other stuff. I haven't seen the product itself. I haven't played around with it. I know that one of our firms in the network is using it and was involved in this beta test. I haven't got anything direct back from them in terms of like what their feedback was,

[00:59:39] but I know that they chose active specifically to kind of replace some of their other systems. And so this is probably part of their longer term plan. Obviously, like I said, there's a whole bunch of stuff missing. So you've got things like FRS 101, IFRS, charities, education settings, trusts, and bits and bobs, all which will have to come down the roadmap. And I would say that's probably a good, probably 12 to 18 months away. If we kind of look at where some other people have been trying to build things like charities, for example, it's taken them at least 12 months, if not longer,

[01:00:08] to kind of get that into the product from the point where they've kind of publicly said they're going to build it. And they have said, actually, now that I've read the press release properly, they have said that they're going to get companies house filing baked directly into the platform as well. So that's, that's coming soon. So like I say, another like cool little extension of what they've already got. I think this is a good place. I think there's lots of opportunity to disrupt the accounts production marketplace because of all the stuff that we talked about earlier on the pod. And then I guess the question then is like,

[01:00:37] what do you do when you're an organization preparing these accounts? And broadly, almost all of those organizations will have a corporation tax return to go with them. Because again, at the moment, there is no good, you know, once you start to get into the complexities of a large FRS one or two client, for example. Did, did you both see, um, doing the rounds on LinkedIn the last seven days, but open AI buying accounting firms. Yeah. Yeah. What, what do you think about that headline when you read it? It's misleading. Yeah, it is misleading,

[01:01:07] isn't it? So, so if I just break it down, so people might've seen this and thought, oh shit, what's going on here? Is this like the end for us? Um, so yeah, the headline was, uh, open AI buying accounting firms. So what, what actually is it, uh, is the open AI doesn't directly buy accounting for, it doesn't directly own and doesn't buy accounting firms. It has, um, taken a, uh, an equity position in thrive holdings, which is, um,

[01:01:39] uh, uh, uh, uh, uh, uh, has been rebranded. So current's been rebranded from creep professionals Alliance, which has bought 48 firms in the U S I think. And this actually goes back to December and so it didn't involve any cash. Essentially open AI are deploying engineers into thrive holdings, uh, portfolio companies for them to better reimagine processes. And as a result,

[01:02:08] in exchange of that open AI taking some equity and they are also getting access to the data in those. Um, I don't think it's just limited to accounting firms, but whatever they are supporting thrive holdings with. So they're getting access to data. So there's a bit of a, probably a bit of a circular economy here. Hey, I'll give you equity. I'll take equity. It will inflate everyone's, everyone's valuations. Um, so yeah, not quite as, uh, impactful, I think as the headline, uh, led out, uh,

[01:02:38] to believe. So I, I think it's interesting nonetheless, because they are, the current are obviously trying to re-imagine, uh, processes. And I think there was a stat around 7,000 tax returns were filed with a 97% accuracy, um, or 98% accuracy and prep time down. So, so they're, they're doing some stuff in there. I think one thing that maybe is quite interesting closer to home is that there's a similar pattern in the multiplier holdings. Uh,

[01:03:07] I don't know if you guys saw, I don't know if you guys saw this back in, it was towards the end of 2025 when on-site accounting founded by Martin Brennan announced a partnership with multiplier holdings. Now there were partnerships always like, Oh, what does partnership mean? But actually I think it was an acquisition. And so multiplier holdings, um, Singapore registered founded by an ex-stripe exec. Um, they've been running, uh, a playbook on professional services and maybe other firms where they are buying them up.

[01:03:37] And again, trying to re and imagine processes or, or, or push down technology or support them in technology, um, transformation. So I think that's quite an interesting one. There's closest to home, but then, um, the, the roll-up model isn't the only one. So we talked a little bit by prosaic. That's just purely general ledger, but this week alone, we saw Nino. If you guys saw this in Amsterdam based AI native accounting firm, or sorry, let's call it a platform, um, raised the six, uh,

[01:04:07] six and a half million euro seats, uh, round, uh, who, uh, they, they said they've got a couple hundred customers, um, and can service 200 clients rather than usual 30 with one accountant. I'm not sure how you kind of work those maths because one client could be like a very small business. One client can be a very big business. So I'm not sure how that actually like compares Apple with apples. But I think, um, the interesting one here is that we're, thrive and current are, uh,

[01:04:36] and multiplier and maybe using existing legacy firms. Nino's not buying firms that starting from scratch. So building the challenger firm from scratch, um, to disrupt kind of, you know, what in their words to disrupt what, what, what, 50 year professional services models always been delivered as. Um, but I think what's interesting is we're seeing three different models converge at our doors or in the profession, which is the full acquisition of roll ups. So the current, uh,

[01:05:06] the thriving current model where, where engineers are going in, we're seeing like a partnership model where, um, multiplier, for example, is buying an existing firm and then deploying technology into it. Um, like for, I guess they did a full acquisition or controlling acquisition. And then we're seeing from scratch AI native challengers like Nino, which I guess at the lower end of the, the price point, someone's just getting technology. And as you move up towards more complex businesses,

[01:05:33] you're getting fully deployed finance partner to support as well, uh, at that price point. So I think that's quite interesting to see how those play out. And then of course you've got the general kind of normal private equity, uh, uh, roll ups in this, in the UK, which are, you know, they're, they're buying firms and they're trying to implement a, uh, technology transformation program and get synergies. Uh, I, I personally made emphasis on the word trying there. Um, so yeah, uh, different, different strategies there. Um, direction of travel, probably similar.

[01:06:03] It's like how can much more be, uh, automated and how can we reduce the human friction, uh, or maybe redeploy the human element to more value adding services. That's unclear yet, but yeah. What, what, what were you guys making of this? Yeah. I was just going to, I'm going to everyone. So I was just going to add one more. More is, or the fully AI practice that doesn't need any humans whatsoever. Yeah. Yeah. And we, which one have you seen in that? Which one have you seen that does that? Um, you know,

[01:06:32] the Alexis Kingsbury who created the, he's not even an accountant and he set up the first fully AI practice. You know, does he have clients? Well, he did. He did. It was Ryan. It was Ryan. It was Ryan just to test it out to see if it was, if it worked and it did. So what's to say that we won't see more of those coming in the future. Yeah. I mean, that, that is like, that, that's something we've been talking about for a while, isn't it? Like the, the risk.

[01:07:02] And, and obviously, you know, I think actually Stuart, um, you're formerly from carbon and I've written about this as well. And, you know, when you look at, when you look at the regulatory environment and for accounting, particularly in the UK, but it is reflected broadly around the world. It's like, you know, actually there's only very limited services that are restricted where a human being has to be involved in a qualified human being has to be involved. Um, and it's certainly not, you know, certainly not tax, certainly not bookkeeping, certainly not broad accounting services. Um, so, so there is,

[01:07:32] there is certainly a future or an opportunity for someone to build that accounting practice, which doesn't have any humans. Um, I think there's still many, many question marks over the success of, of that and, and, and the scalability of that without actually having humans in the loop to kind of review the outputs and stuff. But it's interesting. I mean, I think also to add to like the, the real of things that you, you, you, you smashed out there, Alistair, you know, just literally in the last couple of days in Denmark,

[01:07:58] there's a bit an AI first audit firm launched called a Repodo and they've raised an 8.2 million euros, um, which I thought was really interesting, you know, for a number of reasons. Number one, like in the network, our biggest firm in Denmark is the number one firm out there. And a lot of what they do is, is audit. So they've now got an interesting new competitor on the doorstep aside from the big four. The, um, you know, the, the sort of talk about why they've launched this was sort of saying, look, all it's been stuck in the same processes for 40 odd years,

[01:08:27] manual processing of information, manual verification of documents and all of the peripheral things. And not enough time is being spent on the kind of subjective areas of an audit, which obviously is where all the regulators fall out of this as well, because you know, no one ever criticizes an audit for taking another, another bank statement, another invoice. What they do is criticize them for not looking at the going concern or the, you know, the, the future operating model or the revenue recognition or something like that.

[01:08:54] And so I thought it was really interesting that you talk about that. And yet all of these things, it goes back to what we talked about before when we were talking about this difference between efficiency and customer experience is like, I haven't read anywhere yet with any of these press releases, whether it's the one about your chat GPT and open AI or whether it's, you know, these new firms, not one of them seems to be focusing in on customer good quality, customer experience. They're still looking at internal efficiency. Even the chat GPT stuff is like they're sending engineers and people to

[01:09:22] redesign internal processes to make the practice more efficient and more suitable for the deployment of AI. But they're not looking at like, how do we improve the customer experience using AI so that they get, you know, a nicer, you know, a nicer experience when they do the attachment. So yes, it might, they might have been able to do more volume and it might have been more accurate and it might have been quicker. So that maybe contributes to customer experience. But that's, that's a, that's an outcome derived from, you know, I think it'd be interesting to see what happens with the professions. So,

[01:09:52] so yes, you're right. Some services don't need to be regulated. Yeah. So you wouldn't necessarily need to do that, but are, are the professional bodies going to change their views at all? Can they have a different type of regulation or accreditation for an AI first firm, or even is there a new body that's going to spring out the woodwork? That is accrediting AI firms. And,

[01:10:21] and it's just, it is a different type of accreditation. So, you know, is there an opportunity at a body? And mate, you sounded like AI then. So what'd you say? An AI, an AI firm accredited by AI. That's what you want. Exactly. I had, I had a call with, or speaking to Petri Rantamaki from Visma yesterday. And he, he used the phrase that he heard as I was like, probably, probably Chinese whispers, but we've, we've relayed it amongst a few people.

[01:10:52] Accountants are the blue collar workers of the white collar economy, which I thought was quite an interesting way to put it. As in displacement. Yeah. So, yeah. I mean, I, yeah. I mean, I think that there's still so much up for grabs, you know, isn't there? I mean, that's, that's the thing. And then we, you know, you talk about where else is like professional services being disrupted. Well, you know, here in the, here in the UK, we do have an AI first legal firm that was approved by the SRA,

[01:11:22] where there is much more, you know, granular control and regulation over what firms can and can't do and what their outputs look like. And they won a case as well, fairly quickly off the back of getting that, you're getting that regulatory approval. And I, and I think they said that was the first case one by AI. Now, I don't know the firm. I don't know the details well enough to kind of like say, you know, how much AI was utilized, how much human input was there and how much human supervision was there in that process. But if, if the SRA,

[01:11:51] which typically let's be honest, have been reviewed as kind of like a fairly backward regulator in terms of their approach, although that has been changing in recent years is willing to allow this technology to become part of, part of their environment, you know, that's potentially going to happen somewhere in the future. I think the difference here is like, you know, certainly in the UK is that that regulation is only really relevant for insolvency practice and audit practice. Right. And, and audit, audit still needs someone to sign that audit report at the end of the day.

[01:12:21] And I don't think we're going to do removing that, that, you know, that, that wet signature at any point in time, even if it is digital now. So maybe the, the insolvency practice is a space where you can see more of this because that is certainly an expensive part of the world, you know, from a service point of view, it's an area which was ripe for disruption in terms of like, you're really just processing a ton of information. It's basically a combination of operational business practice, investigative work, and, and, and data analysis,

[01:12:51] all of which we know AI in its broadest sense is really, really good at. And, and, you know, insolvency practices charge, you know, massive premium fees that work because there is ambiguity in the result and getting the result because, you know, not all insolvencies become successful in terms of like the insolvency practitioners are not always paid. I know that, that might not sound right, but that is the way it works, which is part of the reason why they charge premium rates for the work. So when they are successful, they are at least compensating for the work that's lost.

[01:13:21] But maybe that's the area where we'll see big disruption in a regulated part of our world. Let's go. Yeah, nice. Awesome. Well, let's move on to the last story because this one definitely is a little bit more controversial for sure. And this is, this is talking about, you know, something that we kind of touched on a little bit over the last, last few minutes as well is like, you know, who owes the data, right? Who, who has the data that goes into a product and a system in our world?

[01:13:50] And this is kind of all blown up in the last few days, which I've been enjoying. I've got the popcorn out and I've been enjoying this as like Dwayne Jackson, obviously he's building sodium at the moment with brand new practice management solution, challenging the market, doing what Dwayne does. And then apparently you, he's winning a lot of business from, from tax calc and the, and the team over there. And, and then posted about the fact that they changed something in their system, which meant that one of the forms that they had was encrypted,

[01:14:18] something that they were previously able to grab the information from using Jason, you know, which is just a programming language and a structure for data. And, and then the rebuttal back from, from, from tax calc was look, we haven't changed anything other than we're just, protecting our intellectual property around the forms and the structure of that, but the underlying fundamental part of what, what you want as a business, if you're leaving us, you know, which is the customer information, the data that's in that form and stuff is still available to you in a CSV and you can extract it. And, and,

[01:14:48] and this has been bouncing around for a few days. So like, you know, apart from the fact that this is broadly entertaining, um, you know, what are your thoughts on this? I don't understand what the big deal is, to be honest with any software, if you're planning to move from one to another, if, when you're exporting information, they all give it to you in their own format. And you do have to manipulate it to make it ready for the next software that

[01:15:17] you're going to upload it into. So if you can still download it, but you still need to do some manual manipulation, then is that not just what we do anyway, when moving from software to software? Yes, I guess is the broad answer. And I think maybe where this has all kicked off is that they were able to, you know, effectively pull a lot of this information and structure of the data in a sort of, in a bunch of scripts to make life easier. And this,

[01:15:44] this encryption has then made that more problematic by pushing it back into a CSV, which changes the process, which makes it a little bit more manual, but yeah, fundamentally like that's, that's, that's true. Okay. They don't want to work with their competitors, but. Yeah. I mean, you still want the information. I mean, this, this is always the thing about software, right? Is software is about protecting your, your moat in many respects. and this isn't new. I mean, you know, if you had a, you know, if you were on Sage 50 back in the old days,

[01:16:14] the only way you could look at that data was to put it into another version of Sage. Now, obviously technology moved on and we have platforms that move my books, which allow you to kind of grab that data and repurpose it and, and move it along onto another system that, you know, I know from my old world of like doing system implementations, there's still a pen in the backside to get data out of old, older systems and even some cloud systems in a structure that's useful without a lot of data manipulation before you can get it into a new system. So yeah, I mean,

[01:16:43] that's just part of our world. I guess maybe a question for you, Al is like, should that friction exist or should we be looking at moving to some kind of like mandated common data models to allow, you know, transition between systems and things? Yeah. So, so this is actually part of what I was talking to Petri at this moment about, and I, I, I, I don't know enough about sodium and tax calc and the process around it. I've not tried it. I've not, I don't use either of the applications,

[01:17:12] so I can't comment on, on kind of this specifically, but just like generally in the, the, the world we live in, it is a real frustration. How moving data from one system to another, even systems that should talk to one another, right? It's not necessarily lifting your data and using another system. There's a competitor. It's just moving your,

[01:17:38] moving your data freely from one system that should talk to another system is just hard. Right. And it's, it's just like, if we, you know, we talk about customer experience on this, this podcast, we are the customers of those vendors and we don't have a good experience with it. And that's a real frustration for me that vendors don't talk to one another. They lock down data and restrict use. Like we heard from zero six months ago, Ryan,

[01:18:07] you can't use the, or vendors can't use the data to train on, but it's not their data. Right. And so if I engage with another vendor and say, Hey, I want to improve my process or my experience and my delivery by training, um, your model on my historic data, my client's historic data, that's my decision to do it. It's not anyone else's decision. It's between me and my client. And so I think locking down and restricting data, I think that is not in the best interests of, um, firms,

[01:18:37] uh, for clients, our profession as a whole. And it is a poor, um, excuse for not having a proper moat in the, the, the realm that those, uh, vendors are playing in. Right. And, um, I think you can't use that as a poor excuse to hold, uh, to, to, to, to build a moat around it and hold, hold, hold customers to, to ransom. Like you should have a good enough business model,

[01:19:06] a good enough experience, a good enough platform that delivers value without saying, you're not using the data for whatever you want to do. So I think as a profession, we need to grow up a little bit, I think, or sorry, an industry, not professional. The professions grown up, the industry, which is the bigger circle around the profession needs to grow up. And I think we need some sort of standard protocol that makes it easy for data to be moved either from, uh,

[01:19:35] system A to system B that's competitors or from system A to system one, which is a complimentary system. And those protocols just should be some sort of standard that is referenced by all accounting tech vendors. And I think it's just so frustrating. Like so frustrating. Um, so they're, they're my views. Yeah. I guess the question then is like, you know, there is technically like an industry body for software, but they don't really do much here in the UK.

[01:20:05] And we don't, um, you know, we've got multiple industry bodies here in the, in the, in the UK for, from the professional standards and professional services point of view. So like who, who crushed the nail on this, I guess is a question. I mean, maybe, maybe it's done by, um, supply and demand. Um, you know, if, if, if, if, some firms come out with it, then they're,

[01:20:34] they kind of, you know, the cream elevated, the cream rises to the top. And actually Petri was going to take it back to his Visma and his next, uh, quarterly management Visma day in the, the, the topic of, of having something that is more standardized and, you know, seeing if they can, uh, lead the market on it. Um, who knows, but if you, if you build with that in mind, or you have a few people sign up to some sort of common protocol, then maybe you, you start to,

[01:21:04] um, direct the, the, the wider vendors towards that because actually the, the crowd is speaking as in the, us as buyers of the software, because it's easier to use and that's what we want. And so maybe that's the way forward. Uh, or maybe we just all protest at the next, uh, event and, and take picket boards. Yeah. Yeah. I mean, I, I always wonder as well as like,

[01:21:31] is this an issue as we still go through this transition from desktop to cloud, you know, because this is effectively what we're seeing here with this, this particular issue between sodium and tax count, you know, tax count is a product of moving to the cloud, but isn't totally there. I think it has some sort of hybrid cloud setup at the moment. Sodium, obviously fully cloud and built. And, and I guess in, in, in many respects, you know, you could argue that things like APR as an MCP solve some of those challenges that you're talking about Alistair. But again, I also know that soft, you know,

[01:22:00] certain software houses like zero. Again, we talked about them a lot today. They're a classic example where they've got a really good API from a developer point of view, but actually you can't get into all of the endpoints in terms of like the data that you want to access to, you know, you, you, you can build relatively easily if you want to do specific things, but there are certain other things which are just completely off the cards because you can't get our data. Yeah. I think there may be three parts to that. I think if you ask an engineer about zero, having a good API, the developer API, they'll probably tell you it's completely the opposite. Right.

[01:22:31] And so, you know, there's a, there's probably a third of their API, which are, is decent to be able to use. There's a third of the API, which you need to do all sorts of funky stuff with webhooks to get a confirmation of when something's changed, which is like a nightmare to work with. And then there's maybe a third, which just doesn't exist because it's never been exposed. And actually some of the third that does exist, it behaves differently because different departments have built the API endpoints over time and it just behaves differently. Right. Maybe it's not a third, a third, a third, but I'm kind of saying there's an A, a B,

[01:23:00] a C in terms of their, their APIs. So even someone that's even a tool that's supposed to have a good API, there's more modern in the general era of, of kind of tools that we've got is still like lacking availability and access. And just not fun, not, not, not easy to work with. Yeah. Okay. Cool. Well, I think we'll, we'll wrap it up there. And that's a CTA for all the vendors to get in a room.

[01:23:29] Like as accountants, we get in a room and we talk and we share best practice. And we do that for the benefit of building a better firm. And we do that for building a better experience for a client. And maybe the vendors can get in a room and start to talk to each other and build a better ecosystem for us as customers. So there we have it. That finishes off. And we'll be right back to another episode of the DigiTools and the Crew World Podcast brought to you by the Luke. Hope you enjoyed the discussions today. Lots to talk about. We were a little bit zero heavy, as I said at the beginning,

[01:23:58] but hopefully it's useful and helpful to you. As always, we're always really interested to get your feedback and any comments and, and ideas and thoughts on what we've been doing. Also, if you're out there in the tech ecosystem and we've missed an update from you, like come and give us a shout, let us know, and we'll make sure that we cover it properly because there is a whole heap of stuff going on. So it's difficult for us to keep on top of everything. And, and, and of course, look out for all of our events and stuff that's coming up over the next few weeks.

[01:24:25] Hopefully you'll get a chance to come and join us somewhere and we'll, we'll catch up with you.