Who trains the juniors if the bots do the prep? || Be In The Loop

[00:00:00] Hello and welcome to another episode of Digi-Tools In Accrual World, brought to you by The Loop. By now we've already had our major awards event, so we are just covering off the main things that are hitting our radars this week. And after what has been a noisy couple of weeks, this is quite an interesting conversation because we cover all things that are to do with AI, yes again, but we're looking at it from different perspectives. We're looking at displacement and we're looking at how that's going to impact.

[00:00:26] We're looking at some of the new releases that have been made by Airwallex and All Done. We're also covering off some things that we've seen in Intuit and also the AI risks in general that are presented to us. I really liked hearing about some of the work and the way that we can interpret what is happening with Intuit's perplexity partnership and how that can really come through a track of making sure that for practitioners there's a much more sophisticated way of working,

[00:00:53] that's a bit more intuitive and I really, really love John's driving analogy. Not that I really like most of John's analogies, but this was actually very U-rated, I have to say, for you, John. Hey? And before you answer me on that, can you believe we've got the OGs on one episode? For once. Amazing. Well, look, from my end, I'm excited to bring some niche stuff across. Something that, I guess, one of our own has developed. So, yeah, it's going to be interesting to dive into that.

[00:01:22] And just, I guess, in case, can't believe Digital Disruptors won the award this year. Or we were close, we got to the shortlist. But well done to that amazing who won the award. Oh, my God. John, what are you excited about this week? It's so funny. Do you know what? I'm really interested about, you know, I think we're finally turning the corner on the AI debate to be able to have like an actual nuanced, interesting conversation, rather than just either being up in arms or horrified all the time,

[00:01:51] or like super excited but not really seeing the detail or the kind of like the information behind what we're getting excited about to see how that pans out in the real world. So for me, that's what I'm really excited about. And of course, I can talk a little bit about the app that I've been building, but I'm not going to blow smoke up my own ass. So there you go. Well, leave us to do that for you, shall we? Shall we crack on with app news? Let's do it. Are you struggling to attract the right talent?

[00:02:20] Are you needing to shift the busy work? Are you finding you don't have time to speak to clients? Trusted for over 20 years, the Advancetrack Outsourcing Model provides immediate access to accounting and tax experts who understand your local regulations and seamlessly integrate into your team. Digital transformation, robust security measures, top tier talent, no recruitment fees, no wait time. Focus on what truly matters, growing and improving your business and positively impacting the lives

[00:02:49] of your clients. Be more than just number crunchers. Be life changers. I have some breaking news from Airwallex. And I think this was talked about as well in the previous podcast. So really exciting, interesting stuff. Airwallex have launched a separate AI-powered accounting platform for startups. They're calling it T0 or TO. I don't really know. Like, I mean, it's just another word, isn't it? Who knows?

[00:03:17] And we should have given this to Ryan because he's good at mangling these things. But basically, essentially what they're saying is it's built to replace the spreadsheets that people have, particularly when they're in early stage businesses. Obviously, Airwallex has been around for a long time now as a payments platform. And it's also, I don't think, the first platform that's moving into this space. But I guess essentially what they're trying to say is they're not really going after kind of like the broad accounting market. They're trying to sort of target the startup and sort of your founder-led space.

[00:03:46] That's what they're trying to go after. And I'm talking about the fact that, you know, look, this will help you to do some of the bookkeeping, obviously, but then it will help you to build a chart of accounts as your business starts and then grows. And we'll then give you like the important stats and metrics in terms of cash burn, your runway, your margins, and bits and pieces. So I guess they are very, very much like focusing on that niche. I guess interesting to see what kind of reaction you get to this. It is only launching in the US at the moment.

[00:04:14] So the accountants over here, you don't have to get up in arms about it quite yet. But it'll definitely be interesting because obviously the US market has quite a lot of these kind of like AI powered ledgers now knocking about. You know, the adoption rates of those is relatively mixed, certainly in the accounting space, but then the direct market space, it's quite well taken up now. So maybe this is a clever move from Airwallex as a pivot into something new and different. Yeah, I guess it's what they're going after aligns with their client base, their existing client base, right?

[00:04:44] Because they tend to get the new startups, the high growth things where they just want a platform and get on quickly and use and then having the accounting system. So that makes sense. In the same way, the banks have been bringing in accounting ledgers. I think it's becoming a saturated market. There's going to be a lot of losers in this space. We, I don't think we could call any of it at this point, unless you've, unless you've got crystal ball, John, you can see who's going to win or lose. No, no. I mean, there's already a couple of losers on this show, right? So, I mean, who knows?

[00:05:16] I think you're referring to me. No, of course, I wouldn't dare. Yeah. I wouldn't dare. No, I agree, Ryan. I think it's going to become a really saturated market. I guess it's going to be interesting to see whether this niche play for Airwallex works in terms of like, what does the founder market look like in terms of where does it start and stop? And also, who else is going to react to this? Yeah. What are the merits of them doing this? What are the merits of them being the ones to solve for this?

[00:05:44] Because they've been really good at the cross-border payments. They've been really good at card programs across and running those card programs on behalf of other small businesses who have then had to have distributed ledgers, I guess, in some way. What are the other merits then for them taking this type of pivot? Or is it just, okay, we need to innovate somewhere. How can we sell something more to some of these clients? I think for me, it's just this kind of continuous expansion of the footprint of apps in terms

[00:06:14] of they creep into other spaces and other realms. And maybe Wallex are also struggling with their integrations and their ability to tap into the market. You know, there are literally hundreds of general ledgers out there. And, you know, some of them are easy to integrate into and some are not. So maybe it's better spending their time building their own footprint rather than trying to continue to break into that existing space. Why are they at pains to then make out that it's, okay, this is not AI, this is deterministic. And they're very clear about that.

[00:06:41] But are we speculating that this is just also all some other thing? Yes, sir. They've just been listening to Dave Selleck a lot. It's like deterministic is the word we now need to use to convince people. So that's the one we're going to go down. I mean, my take on it is I think they crave a bit of risk for themselves here in that they were becoming kind of embedded partners for certain apps.

[00:07:08] And as apps are now expanding, you are going to compete. And we've seen this like at Zerocon, right? Zerocon's expanding. Dex and April and Max are now really directly competing because of the way they're changing the pricing structure, et cetera. I think they're just everyone. It's just going to be messy, right? And there was there's a really good LinkedIn post by Trent McLaren the other day where he talked about it's going to become like the Wild West. Everyone is just going after everything at this point. It's so easy to build.

[00:07:36] And that I think normally happens when there's disruption. And then the ones that actually work start to gain traction. They start to get word of mouth. People start to go towards them and then everything else falls away. And I think that's just the space we're in. I think everyone's trying to get a piece of everything and it's not going to work. Drive something, Kelsey. I think it's going to get people talking. So Ainsley Damery has launched All Done and the claim is punchy by him as well.

[00:08:03] It's a full set of UK limited company accounts in one hour or less. And that includes lead schedules, the working papers, analytical reviews, reconciliations, the corporation tax, the computations on the completion sections and also the points for partner and client care. So it's a very big claim. I think it's worth covering because Ainsley understands the accounting market and spent quite a few years training firms on how to package and sell service lines.

[00:08:31] So he knows exactly where the pain sits. He's been there alongside the accountants and also things like the production work eating into the margin. So I think the positioning on this is super deliberate and they've jumped on this sort of All Done is only for accountants. There's no direct campaigns to business owners. I wonder why he's saying that in amongst the last. Yet. Yet is always an important thing to add to the end of the sentence.

[00:08:59] So it's a clear response to the platforms that have been marketing more aggressively, the end clients. And we're going to name known names. But from my side, looking at this on the, you know, in terms of what happens if the production layer really does collapse from 15 hours to one, I don't really understand there how you, the changes that are made, they're saying they don't remove the accountant, but the accountant

[00:09:26] is closer to the review, the judgment, the sign off and all the client bit. And, you know, it's made it clear that these numbers are not generated by AI because everybody seems to want to both lean into AI and lean out from AI at the same time. But they're calculated by its engine from the UK tax legislation and accounting standards. So AI is only used for things like wording or the mappings and options. And then the accountant reviews that decision.

[00:09:55] So they make a really big distinction because if you're producing something that's like close enough, then obviously there's a bit of a risk in there. So, and they've also said they're only supporting UK small limited companies, micro entities. It applies to companies at 2006 CTA to, how do I say this? 2010? Yeah, 2010. And current tax rates, including the employer NIC and marginal relief. So there's no integrations. It's not going to link to HMRC. You've got to use some other tool to then make the submission.

[00:10:25] You've got everything there to start with the production, improve the working papers, improve the calculations. And they've also got a cold review challenge, which is super clever, where you can put 10 jobs you already have signed off through the system and see what it comes back with. And I think that's a much better way of testing the software than a demo. So I'm kind of going back to your point on this, John, is like so many AI stories are out there. But this one seems like a really good one for firms that are still under pressure

[00:10:55] from MTT and then also are being squeezed by the clients maybe who still don't send the records properly. And they're able to get a much better capacity story. What do we think about this one? So I've watched the video and the demonstration that Ainsley posted with this. It's incredibly long for something that should take such a short amount of time.

[00:11:21] It's the most boring video I've ever watched in my life for a product demo point of view. It could be about 25 minutes shorter than it actually is. I think there's some issues here. It doesn't integrate with anything. So it's a waste of time, you know, because if it doesn't integrate into accounts production systems or doesn't integrate into HMRC or Companies House, what's the point in using this system as a standalone? If I have an ideal client where I've got all of the information and everything ready to go, I can do all of this in an hour easy because accounts production and the computations

[00:11:50] and stuff are not difficult with existing systems. And I think when you look at, when you watch the demo video, it is a perfect client where you're not having to go back and forth and chase information and ask queries and clarify on things. That is where the time is spent when you're in the working papers. It's not, it's not time spent on kind of like you're drawing up analysis and you're understanding the implications unless of course it is quite a complex organization, but they're not, they're not targeting complex organizations. They're targeting micros and small companies.

[00:12:20] So it's a great idea. I think we're going to see more and more of this in terms of like where AI can be spent on things and the firms are already doing this. And obviously we've got products out there, you're helping and working around some of this already. But for me, this fell very flat. Why is the demo so long? I don't know. It just goes, it walks through stuff so slowly. It's unbelievable. And, and, you know, it's like one of those where it could be dialed up to like 10 times speed to, to, to get through it. Can you do that though? Can you put it on 10 times speed? No.

[00:12:50] Oh, it's not in, oh, okay. Well, I mean, I agree with most of what you said there, John. The one, the one other thing I think that's worth pointing out is that I've found historically with accountancy firms, the, one of the main things they want consistency on and everyone to be doing in the same way as working papers. So to have it where you can only do it in this small niche, they're going to look at it and go, well, they're going to have to use a different tool for these complex clients. And it's very rare that you've got an accountancy firm that only deals with those small ones. It does happen, but it's rare.

[00:13:19] So it's a very niche area that he'd be going after, or he needs to broaden it. Now, it's a good starting point to say that. It's a good starting point apart from possibly the demo video, but it's a good starting point in that we try. It's an area that needs to be resolved. It's an area where we're going to see more focus on and getting something out there and demonstrating it can work is good. But I agree. Can I probe you on that? So are you answering that with the lens of the mid-tier accounting firm that is looking at multiple different clients, different sizes?

[00:13:45] Would this be something that is for all those that are working with only small and micro entities that maybe are one man bands and want to then have something that's a little bit more intelligent than like your average Grok bot that this could then handle it? You're smiling, John. I reckon that's, yeah. Yeah, I mean, this is a pitch for the micro practices or the single platform practices broadly, for sure. But, and there's a lot of them, right?

[00:14:13] I mean, it's not a small market, but Ryan's right. And as soon as you get some level of complexity, you kind of, you know, you're done with this product, unfortunately. Well, I've worked in those small firms. Yeah, they might have 90% that are simple clients, but then you've got 10% that's still difficult, challenging, painful, which you need the tool to be able to handle. And that's my thing is that once it gets away from that and it can't handle it, now you're looking at two systems. And that's usually just an instant rejection.

[00:14:45] All right. Well, let's pivot away from something that can do everything into something that's very niche, something that's looking at the problems that no one solves. And we were banging on about this for years. Well, finally, someone has. Someone with an amazing vision has solved this. And that is with the release of One Sixth, which hopefully I've said the name right. I can tell everyone what's saying names. You, the dry ice, right? I feel like you're building to a crescendo. I can hear like that, you know, I need a hero, dark, dry ice. It's like, come on, add some extra.

[00:15:15] Yeah. Okay. Sorry. Well, you've thrown me off now. So I've gone off track, but essentially in that very, very niche area of VAT pass exemption and margin schemes, which no software seems to be able to handle, right? You have to extract everything out into Excel, manually manipulate stuff and then put it in. Even HMRC MTD rules said no one's going to be able to resolve this. So it's just the exemption, right? But most businesses need that solution, right?

[00:15:43] I've had to create a horribly complex Excel thing that the file size is so, so big to try and do this that actually on some clients' machines, it doesn't even work. Well, we now have a tool full and I'm grateful that this has been released. I'm going to be tapping this up. I just have to find a time when this person that's created is free, which is very hard to find because it is from the infamous John Toon, isn't it, John? You've been working on this tool. I have. Yes. Thanks, Ryan. Yeah, I've been working on this for a while.

[00:16:13] It started life as a little Chrome add-in that I coded up myself a long time ago that used to sit just on top of Xero when you had a motor business and you wanted to figure out the margin calcs and post that back into Xero. And yeah, now it's evolved into a fully formed app and no longer a Chrome extension. So we've got to get done away with that. But yeah, it's something I've been working on for a while, like on and off, as you can imagine.

[00:16:37] And yeah, I just know that there's a big niche and a big opportunity there in terms of, you know, there's lots of businesses that are affected by margin schemes, partial exemption, and sometimes a combination of both. And it's not to say that I want to replace those complicated spreadsheets and stuff, but I know that when you kind of do anything broadly spreadsheet related, unless you're using like one of the few integrations that's out there with the GL systems, is they're kind of like fixed in time.

[00:17:01] And so if anything gets posted or changed or anything else, then it often doesn't get picked up and captured. So yeah, so 1.6 is out there. It's in beta at the moment, predominantly across most of the schemes and most of the integrations. So it does integrate with all of the major cloud platforms. And the reason we're in beta is just, I want to just make sure that we iron out any of the kinks and, you know, just get some real world user feedback aside from just me playing around with it and seeing how it works. And yeah, we'll be pretty responsive to that.

[00:17:30] So we do already have, I think, about 10 beta testers signed up. So we're working with them now on like how those integrations with Xero, QuickBooks, et cetera, work in the workflows. You've got a couple of questions though, John, on this. So I've got a question that says Irish cash accounting, VAT, question mark? Yes. Yeah. So that's a scheme that we are going to build at the moment. It's not there in the product at the moment.

[00:17:55] But yeah, so the weird thing about Irish cash accounting is it's different from UK cash accounting where we do, you know, if you move from a cruel account to cash accounting, you do the cash accounting on both sides. So on sales and purchases. But in Ireland, they don't do that. You stick with the cruels accounting on the purchases side, but you move to cash accounting on the sales side. So you've got that immediate like ad debt relief on the sales side, but you actually get a cash flow advantage because you're still claiming the VAT back on your purchases, even if you're not paid for them.

[00:18:24] So it's slightly complicated because none of the platforms handle that because broadly they'll only accommodate the UK version of cash accounting. There is a product in Ireland that kind of works with Xero on this already. It looks a little bit clunky, a bit nasty. So mine's hopefully nicer and better and across more platforms. So it should work more effectively. And you've got it in beta. So are the software users at the moment, are they paying for 1.6?

[00:18:53] So at the moment, all the beta testers are free. So they're getting to use it with their clients and with their live data and testing it out. And so like one of the testers we just had go through it found, I think it was a golf club that they were testing and they found that the historic partial exemption calcs had been done wrong. And so that got picked up straight away. And unfortunately, it was wrong to the detriment of the client. So then that prompted a slightly difficult conversation. But this is a client that they picked up from someone else.

[00:19:23] And so they weren't responsible for it, luckily. But yeah, it's one of those things. But I think broadly, that's one of those challenges sometimes is that you do have people using these schemes and not necessarily understanding the rules. And I'm not saying that the platform will fix all of those ills. But we do try and make it sort of relatively codified. Yeah. And we can say this isn't a lot of pony is going live, but it does handle the equestrian margin scheme. So it's getting pretty niche in what it works with.

[00:19:53] But is this aimed at accountancy firms or end users, John? Are you bypassing the accountants and going straight to end users for this? Yeah, the account world's dead to me. No, we can be used by both. Because clearly, there's lots of accountants who are doing this kind of work on behalf of their clients. Lots of accountants and bookkeepers, I should say. But also, there are lots of businesses who have these challenges about their own finance teams.

[00:20:19] And so there is the usual kind of like SaaS pricing. So if you're an accounting bookkeeping firm and you want to scale up, then you'll get some significant discounts. If you're an organization, then the pricing is slightly different. But there is some scale mixed in with that. And then there is also some charity pricing. Because again, charities quite often are caught by things like partial exemption and some other complications. And there can be multi-entity as well. So we kind of have the ability to capture some of that.

[00:20:47] And we have integrations with the charity commission and other places. So you can just immediately verify your credentials. And then you'll immediately qualify for that discount as well. Cool. Well, yeah. We need to have a discussion because I've got some clients for this. But this isn't a SaaS area for you, John. So let's move on. Who's got another news to it? Me, actually, funnily enough. So let's move into the audit space away from taxation.

[00:21:15] So AuditBot is a product that, if I remember rightly, spun out of one of the larger accounting firms, the guys behind it, Jack and Jared and others. I'm trying to think. With the RSM, I could get this wrong, but it's one of the larger firms anyway. But AuditBot has been knocking about now for probably a couple of years. And they've just launched AccountSpot, which is their answer to the financial statement review process.

[00:21:38] And they've connected up with the HAT group, who some people might be familiar with because they provide audit methodologies and checklists and some of the other stuff. Obviously, there are other choices on the market if you wanted to go down that route as well. And so you can now do things like disclosure checklist reviews in Audit. We're using AccountSpot, which is a massive area of time-saving opportunity.

[00:22:01] Because for anyone that has ever done an account checklist or disclosure checklist, you're looking at something that starts with about probably 100 to 300 questions. And then broadly, depending on the complexity, you can expand that to sort of 500, 600, 700 questions on a massive checklist as you go through the financial statements. And most people that fill them out basically take last year's and they just copy the boxes, yes, no, not applicable, and just hope that nothing actually fundamentally has changed because we just turn into robots when we do this kind of work.

[00:22:31] So I think handing that work over to a robot to do on your behalf is a very clever idea. They're not the first to market. There's a couple of other people already doing this in the market. Data Sniffer, in particular, have been doing this since the tail end of last year. But this is kind of like the way that we're going to see. And I think, you know, when we're talking about where AI can have a significant impact, this is one of those areas where, you know, having a platform structured like this really, really helps.

[00:22:57] And I can say that from my own personal experience, because we tried to build this when I was at Beaver & Struthers using ChatGPT. And we could never quite close the gap. We could get to about 70% to 80% reliable, but then the kind of like that missing tail was always like just unreliable. And that was broadly down to like how you manage the context of such a large checklist versus quite complex financial statements, particularly if you're doing it on your big like sort of your 30, 40 plus page financial statements.

[00:23:24] It just started to get a little bit wild at the tail end and not necessarily hallucinating, but just getting a little bit lost in terms of like not quite nailing down everything. So it's just unreliable. Again, like I say, there's a fair bit of competition in the market for this, but you know, this would be an area that if you're looking to save significant time and it's probably at the sort of senior level, manager, director, partner level. That's a huge time and cost saving for an organization. Yes.

[00:23:50] And it's always good to hear advances, not only in the account side, in the audit side as well. I've got something that's not so, I don't know, dramatic. It's just kind of more news that we've seen over and over again. Intuit and Perplexity have announced a strategic partnership, essentially bringing the QuickBooks Online and MailChimp information directly into the Perplexity system.

[00:24:12] Now, the main thing they've done around this is actually QuickBooks or Intuit have expanded their MCP server to enable the cause of the interrogation that would be required from the Perplexity engine. So there has been some work. It's not just one of those marketing announcements. And they have put some work in to enable the interrogation that anyone that's on Perplexity, that's a user Perplexity, would need to get the information out of both the QuickBooks Online and MailChimp platforms. So this isn't the first.

[00:24:41] They've done other announcements, but they're broadening. So that whole thing of every single finance ledger, or the main ones, connecting to every single AI platform is just continuing. I find this really interesting, right, because Perplexity is broadly a research model. It's like a technical scientific research model. That is what it's effectively pitched to the market as. So I find it interesting that Intuit have gone down this route in terms of like an MCP. Like you say, they've already done this with others.

[00:25:11] So it makes sense to kind of continue to expand that out. But for me, if I was Intuit or if I was Perplexity, I'd be looking at this saying, hey, you know, look, we know that you guys operate in a really technical area, whether it's accounting standards or tax or whatever else. And that is kind of like broadly connected to what you're doing. So the advantage of connecting to Perplexity is that we can give you like validated, reliable outcomes when you're asking technical questions about what's going on in the GL.

[00:25:37] So I don't really understand why they haven't like lent more into that, because for me, you know, that is where you could differentiate from an integration with the likes of Claude or ChatGPT or someone else, because because of the technical nature of Perplexity and the way that it's been built to be able to give you the sources of information to validate those and to make sure that they're reliable in a way that the other models don't do that.

[00:25:59] Well, something that they have specifically focused on in their announcement is that they basically said that, and I won't say this exactly verbatim, but those users that use both platforms now will be able to put the expertise and intelligence to work from their Perplexity computer going from research and answers to securely executing actions in one seamless flow. So they're basically going, because it's reliable, because it's giving you all the background, you can now act.

[00:26:26] You can put that into actually doing things inside the Intuit platform. So I think that's the reason for it. Exactly the reason you said, right, because it's given more trust. It's given more trust back to those in the end, just to then make decisions rather than just putting out and having, which I guess a lot of the tools are focusing on is that kind of insights rather than actually making actions inside the platform. That's the reason they've gone down to, actually. Do you think that then blows out the water all of these tools that we've seen fairly recently that are sort of saying,

[00:26:53] hey, we'll help you to validate like VAT invoices and the treatment of those? Because this is perfect for Perplexity, right? You have that connection in there and you could just say, hey, lurk in the background of the system and just check every transaction against sources of information and flag something that doesn't look right. Well, on that, look, Bill Gates has published his essay and he does talk about the accounting worker who is replaced by, by a bot.

[00:27:19] And he's picked accountancy as his primary sector for the illustration of AI replacement. And he's written a whole essay on that. It's exactly what he said. He's picked a good profession because accounting is a white collar displacement. We've covered a little bit of that in the last episode. But the bottom line of the firm is very exposed right now because, you know, automation isn't targeting the partner level relationships that we're talking about here.

[00:27:46] What you're talking about is those foundational tasks that are initial preps, the early reviews, lurkers. There's things on the TVs. Of course, we call them the lurkers. It just makes them sound very creepy. But looking at things like the preliminary tax computations and the tidy ups for the files.

[00:28:05] I think from a business perspective, when a practice cuts 10 hours off a job, that profitability jumps up immediately when you've got something that's trying to catch things as it happens. And the technology will help then render a practice much more lucrative, which is great for the partner level. But I guess like the question I really have around this is a catch 22.

[00:28:31] The financial reports will reflect the short term gains. But the instincts that the junior level won't, it just won't have it because it won't have been developed. It feels like a bittersweet pill for those new accountants that come up the track. I guess for you both, I mean, you know, what do you think about have firms already compromised that junior development anyway? Or is that overly critical?

[00:28:59] Because if the trainees skip basic preparation, how do they get the expertise needed to make sure that that thing that you're saying is the lurker is actually doing the right thing? Yeah. Yeah. There's, you know what? There is completely different viewpoints out there. There are some people that are going, the model is now, you've got nothing at the bottom. You've got a wide middle management that are doing all the reviews and then you've got the top end. And others are going, well, they're seeing something completely different. You've got a huge bottom layer that's kind of building this stuff and understands it and really effective.

[00:29:27] And then a recondensed middle management layer because the quality of the outputs are now higher. So there are, no, no one has got this solved. I think the bit that I would say is we can't ignore the junior level. All that is the same way of outsourcing everything. We just create our own problem inside the UK. We need to still develop people. And that hybrid role of understanding technology and accounting, I think it's going to become more and more critical as we go on.

[00:29:54] I was going to bring up something a bit later, but I want to bring it now because it's very relevant for what we've just been discussing. And that's that Michael Gould, or Gould, sorry, co-founder and former CTO of the business planning platform, Anna Plan, if I've said that correctly, has basically been flagging that there's a heightened risk going from in the past where we create everything on that big spreadsheet, right? Where all the financial planning was on that spreadsheet someone created 20 years ago that no one could touch because if something breaks, we can't do anything anymore.

[00:30:25] To now relying fully on AI models that no one can explain where that information has come from. So on that, they're going to rely on it. They go, oh, I've got this information. Someone then asks them a question, interrogates it, like why is that? And they haven't got an answer for them anymore. So at least they could with the Excel. It was just more risky of it falling over. Now they're going, oh, I've got this brilliant answer, but we can't explain anything. We don't know actually how relevant it is.

[00:30:50] And so all we're doing is that whole mistake we had of relying on spreadsheets is now pivoting into relying on an engine that no one understands fully. So I think this perplexity thing is great. I think that's the other thing with the Bill Gates side is actually until we've resolved that and it's all fully audit trail, it's all fully defensible, then we've not really resolved anything. We're still dabbling. So I just want to bring that in now because I thought that was relevant. Sorry, John, if I cut over your opinion on this. No, no, no. I think it's really interesting, actually.

[00:31:18] And I mean, I'm sure you and I, Ryan, at some point have inherited complicated spreadsheets from somebody else in a firm and then spend ages trying to understand how that number or that column of data or whatever it is you're looking at arrives at that point, you know, and where all the information comes from. Particularly if it's had things like macros or something else going on in the background as well. And, you know, maybe that's the closest I can get to like where we are in the world. And I think you're absolutely right.

[00:31:46] There is lots of talk amongst regulators, whether it's the FRC or the ITAW or others, you know, around, you know, how do we kind of interpret the outputs of AI and how do we understand kind of like what's going on in that black box? And in some respects, I think that's almost like beyond our capabilities. Like I'm not an AI expert to the point where I can tell you like how has Claude or ChatGPT got to an outcome and a response. I can kind of hypothesize and I can kind of figure out much of it.

[00:32:15] But I can't get you. I can't do the full technical analysis. I think going back to like the junior like trainee problem, right? Like every accounting firm for the next 100 years is probably still going to have to have some trainees and some people coming through to kind of keep refreshing the employee pool to keep pushing the boundaries to have that throughput of people going, you know, able to like stick with a business and an organization as they grow and develop, become managers, become directors, become partners, et cetera.

[00:32:45] I just think that, you know, the problem is, is like we just we just have our old school thinking caps on when it comes to how do we train people? Like we're still we're still wedded to the fact that doing double entry, ticking off invoices in order, you know, filling out your tax computations teaches us something about the job. Fundamentally, I can absolutely guarantee that I never learned, you know, anything from double entry bookkeeping other than the mechanics.

[00:33:10] I never learned anything from ticking off an invoice when I was doing an audit that I couldn't have learned some other way. I didn't learn anything from filling out a tax computation because I could have just been given the numbers without understanding how they were calculated, you know. And so, you know, we you know, we need to be thinking a bit more, you know, in a more sophisticated way about how we go about training.

[00:33:32] And, you know, I don't normally do this, but I'm going to tip my hat towards where the big four are leaning into this because they're looking at their training and development of their trainees going forward and basically saying, look, instead of being a trainee for the next three years and getting through your ACA, et cetera, et cetera, you're basically going to become a manager in one year. That's what we expect now. You know, we're going to try and give you the level of experience and the knowledge and information that would normally have been gathered in four to five years of practice life.

[00:33:59] And we're going to try and push that through to you in a way that means that you can get to that level within one year. And for me, fundamentally, that is where we have to get to. And that kind of leads back into that kind of like that diamond shapes kind of like shape of the firm of the future, if you like, in terms of having a lot of people somewhere in the middle and people at the top. And I don't know if it's right, you know, but again, the one final thing I will say is when I come back to your analogies, right, I really like a car analogy.

[00:34:29] And so like as drivers of vehicles, right, over the last hundred years, we've gone from having a vehicle where there was a bloke walking in front of it with waving a red flag. We no longer have that. So that's a skill and something that's completely evaporated from driving. We now have like a synchromesh gearbox. Now, actually, if you get an electric car, you don't have a gearbox full stop.

[00:34:49] So, you know, there are a whole bunch of things that we don't learn as drivers that people previous to us have had to learn as a consequence of like how does a vehicle run and operate? And yeah, at the same time, the purpose of becoming a good driver is to get from A to B safely without endangering yourself, your passengers and other people on the road. And it's to get there as efficiently, as effectively as possible. I think when it comes to accounting and audit, bookkeeping and tax, like we're still trying to achieve that journey, right?

[00:35:19] We're still trying to get from, you know, a reliable audit, a reliable set of accounts, a reliable tax return. We're still trying to get from A to B. But the method that we get there is just slightly different. You're absolutely right. What you're saying sounds a little bit akin to what Sir Kindersen Cassidy was sort of pointing out at Xerocon, which is the fact that, yes, OK, you could use that time to hone your professional acumen and entangle lots of mess and see the errors.

[00:35:47] Or you don't need to know how the pistons fire in the engine and what's wrong with the gearbox. Just use the vehicle as if it works. And there's more sophistication in your experience and what your job is. And we've got a lot further forward with it. So I sort of feel that on a personal level, yes, there's structural vulnerability in just about every technology that we use.

[00:36:12] But we don't understand all the nuance between the technologies that we create and we use for ourselves anyway. So, yeah, it's just that I get excited hearing about that because, you know, you hear about people going into the world of work and thinking what's going to be there for them. Well, what's going to be there for them is the benefit of five, six, almost 10 years of what would have been work experience jammed into one year. They're going to be super fast, these guys.

[00:36:40] And I think that's what's exciting for me, right, is, you know, again, when you, I think for most people, when they think about their early stages of their accounting career in terms of like that training experience, whether it's the big firm or a little firm is like, yeah, it was fine. But there was so much of it that was just, you know, manual data entry. It was robotic, repeatable work that didn't really add a huge amount of value. You could almost argue that quite often it wasn't really that interesting and engaging.

[00:37:09] And, you know, at least for me, when I reflect on my career, you know, the bits that the most exciting bits were the bits where you effectively were getting to meet the clients and speak to them and engage in some kind of conversation about what's happening with their business or their organization. You know, whether that's strategy, whether that's personal stuff, whether it's broadly just talking about the weather or whatever it might be, you know, it was just those kind of things.

[00:37:33] And so, you know, and again, you know, lots of firms do this in different ways, but again, traditionally, you know, the trainees are kept away from those conversations in the early parts of their career. And it's only as you kind of move through the ladder that you get more experience. And maybe this is just the way it's going to work, you know, and that's the challenge, I guess, for management, whether it's partners or directors, managers or whatever. It's like, how do you how do you adapt to that requirement?

[00:37:57] You know, in terms of like we're in this very strange transitory phase of firms broadly are very busy, still got lots of lots of the manual work to do. You still got lots of the review work to do. You still got lots of other things that take up your time. And yet at the same time, we're also saying, oh, do you know what? The way that you train your junior employees is now literally about to change. And you're probably going to have to mentor them a little bit more directly. You're probably going to have to take them out to more meetings on a more regular basis.

[00:38:26] You're probably going to have to get them to engage with clients in a different way. All of that takes time and effort and other things. And where's the bandwidth, right? Well, they're going to have to sign up to John Toon School of Etiquette, aren't they? It's going to be the one that I will be putting my name down for. Anything else this week that was interesting? I think I've got one more. So I have one more as well. Who's going to go? Should I go first? Gentleman first. I'll go first.

[00:38:56] So we picked up a report from the Startup Coalition who have got input from a number of organizations and stuff, including the Entrepreneurs Network, Enterprise Nation, Xero contributed to this, the Center of Finance, Innovation and Technology, City of London Corporation, and a bunch of other things. And then this was actually pushed out in Politico. So a very highbrow, right? And they were talking about a number of different things. So I'll kind of cover those, and then I'll dive into the one that I think is the most important topic.

[00:39:24] And it's certainly something that appeals to you and I, Indy. But the first kind of points that we were talking about was that apparently, and I don't know how they validated these numbers, and I'll have to dive back into the report to figure this out. But they reckon that British businesses spend 11 million hours a year proving they are who they are when they're engaging with the UK government.

[00:39:44] So I'm guessing that's not just HMRC, maybe Companies House and other things, but it's a bunch of other organizations in the mix as well, which is like, you know, that's absolutely like staggering, isn't it? It's a mind-blowing set of numbers. Only 37% of businesses say they spend too much time proving who they are, yet three in five have had something go wrong because of it.

[00:40:05] So either a bank account opening being delayed, access to finance being held up, or even potentially losing contracts, which again is like, you know, not what you want to sort of like experience, you know, in an economy, especially one that is kind of like going through this digital transformation at the moment. Basically, they were saying that every check takes approximately 20 minutes. So this is part of the reason why people don't complain because they're kind of like, you know, it's not consuming so much time that it's got really frustrating.

[00:40:33] But the problem is, is like quite often if it racks up, you know, if you're having to do this with multiple, you know, multiple government departments or whatever, then obviously that 20 minutes suddenly like becomes a huge, huge amount of your cost and things. And so it's a little bit crazy. And then another interesting stat was that a lot of businesses don't actually know what tax they're paying when and how much, which again, just blows my mind as an accountant working with businesses for 20 odd years.

[00:41:01] Like that, that kind of like fries my brain a little bit. And, and so, you know, it's, it's a little bit, a little bit crazy, but also they're effectively saying that as a consequence of like all of this, your friction in the system, they're kind of like calling this a tax on business. You know, I guess it's the classic red tape thing. Right. So, so the answer from the report is to talk about what, what they're, what they're trying to do is like, say, look, you know, rather than having all of these different things where we've got things like your directors and PSCs having to identify themselves separately to companies house at the moment.

[00:41:31] And we've also got like, you know, NHS numbers, we've got a whole bunch of other things. It's like, why don't we have like a digital ID? So leaning back into this, you know, idea that's been rumbling around in politics for at least 20 years in the, in the UK. And obviously kind of comes and goes in waves because people get, you know, people get worried about, you know, their, their ID and their personal information, even though when you open your wallet, your purse, your Google, your wallet or your Apple wallet, whatever, you've got like thousands of cards in there. You've got thousands of IDs.

[00:42:00] You're giving away pretty much all of your personal data without even, you know, giving a second thought to it because it's convenient. And yet when the government is trying to make things convenient, everyone gets really upset and annoyed about it. So there's some slight, you're moronic kind of like thinking in there. But effectively what they were trying to push down is like this kind of concept of a biz pass, which they're saying is not a, it's not a digital personal ID. This is something specifically for businesses and organizations that have to interact with these things. But for me, like those are one and the same thing.

[00:42:27] And actually I think having kind of like these things integrated and combined would be the way to go. But Indy, I know that you and I are super keen about this because we've spoken to some government officials about this. We've spoken to some HMR people, HMRC people about this as well. So we see them feel the frustration in our space as well. And we spoke to the Department for Trade and Business as well on this. I think you're right.

[00:42:52] I think that it makes sense that we do it if it's a convenience and that this would be a convenience in that case. Because, yeah, perhaps we don't want to enter 1984 or some sort of George Orwell novel and be watched and surveilled for everything. But this is one of those things that costs a lot of time and money. And you pointed out quite a few things to the Startup Coalition as well.

[00:43:17] So, and it's really, it's really great to see that they are now taking this more seriously because they've allocated, you know, an agenda for this. And they've got a team looking at it. So hopefully this should just stop that point of, you know, creating onboarding friction. And also what it does show, though, is that the accounts play a role in this, a really critical role.

[00:43:46] That if you want to validate those businesses and you want to make sure that those numbers and that business and that individual and all of this is like above board, the accountant is a pivotal role and still not protected. Which for me is why the next logical thing should be a part of one and the same conversation in a way. Because it's for the, you know, it's for everyone's benefit. It's for the business's benefit because it gives them some form of protection too. And I think we all just know where we stand. So I'm quite excited for this.

[00:44:16] I think you're right. I mean, accountants, bookkeepers play a huge role in this. And I think, you know, we could either be massively supportive of this or we could kind of kill this in its tracks. And, you know, as a, you know, having worked in practice, you know, for so long, I know that that process of like, you know, client onboarding, for example, doing the due diligence and everything else. That even without all the technology that we've got and the great tools around AML and other bits and pieces, it's still a manual process. It's still frustrating.

[00:44:44] What's most frustrating as well is like where you've got a client that maybe has engaged a lawyer and then gets handed over to us or vice versa. Is we still got to go through the same processes and duplicate in duplicate, even though, you know, we should in theory both be relatively reliable in terms of doing that due diligence and getting it, getting it right.

[00:45:01] Right. So, so I think you, along with this kind of concept of your digital IDs and other things, we've also got to make sure that the quality of what's going through these systems in terms of like, who is responsible for what, how these, some of these functions are performed. Is as, as, as, as high quality as possible. But, I mean, the big thing that I want to avoid is this kind of like, you know, like you say, the 1984 kind of George Orwell thing, right?

[00:45:24] Like we, we give personally as individuals, the vast majority of us, apart from the weirdos with the aluminum foil hats, you know, we give away so much of our personal data without even giving it a second thought when we apply for a credit card or whatever else it is. And, and yet we just, for some reason, there are certain people who just don't want to do that with government, even though in theory they should be the most trustworthy source of like keeping that information.

[00:45:49] And on top of that, it would, there are circumstances where it could almost be in your favor from a security perspective that your business is protected because we have this single business identifier and that therefore it almost feels like it's more encrypted than how someone could take advantage of another person's identity. Right now, we saw that with some of the business owners that have had fraud occur at a bank level.

[00:46:17] Yeah. Yeah. We have, I mean, I mean, company's house has the proof scheme, which allows you to like put some restrictions around what you can file at company's house and the method of that being filed. So, so it removes the ability for someone to file a paper form, which is where quite a lot of fraud can happen. But it also means that you have, you're monitoring over your records in a way that doesn't happen. You're out of the box. Problem is, it's like most people don't switch on. They don't know about it. It's not there by default. It's almost one of these things. You know, I, I broadly, when it comes to a lot of these things,

[00:46:45] I just think these things should be done by default. I don't think there should be an opt in. It should be, it's a bit like you, where we've moved to with like donating your organs, right? It's an opt out process for me. Almost everything in life should be an opt out because you should have to make a, you know, a thought out and, you know, and, and fixed decision to choose not to do something rather than to choose to do something. Because that unfortunately is just the way that our society and most people I think seem to work. Right. Right.

[00:47:15] Well, I'm going to end on a completely different story, which takes us away from all of that, but it is with Join and Join who have sponsored this podcast. And they have a new forecasting tool, which lies in the functionality. Some, some very clever things around maintaining the full integration across the profit and loss statement, balance sheet and cash flow. So it's a full three way. Most practices are accustomed to a far more fragmented workflow. I know, I know, John. Hold that for John.

[00:47:44] So extracting figures into custom spreadsheets, manually reconstructing projections, altering a single variable and watching the balance sheet detached from actual liquidity while multi-entity setups collapse. Join is aiming to pull practitioners away from what is the offline workbooks and embed that logic directly into their central reporting suite. So you begin with things like your historical figures.

[00:48:13] You define that baseline, whether that's through the trends, through the expansion, the blank setup or automated suggestions. And then you adjust parameters while the calculations flow across all three schedules automatically. So it's really good on operational efficiency in terms of what it promises. And the intelligence layer operates within those strict guardrails.

[00:48:35] So the practitioners get to outline the target scenarios, such as something like growth spikes or revised overheads or recurring seasonal shifts that happen when, I don't know if it's like someone that's cutting down trees and they only work in the summer or the system translates those inputs into logic rules that remain fully reviewable and adjustable. And it also is an important operational shift because it gets to look at it from the resource delivery perspective as well.

[00:49:01] So, yeah, they've stripped away a lot of mechanical friction points and the focus for the delivery team is really on moving away from what is like the tedious maintenance of things and doing things that are a bit more sophisticated. It's kind of what we were saying on this podcast about, you know, the shift from juniors doing something that might have been just like the basic level they don't really learn a lot from.

[00:49:25] So it's a lot less energy wasted in repairing formulas and much more in the validating the core logic behind why the business is doing what it's doing. And I think the multi-entity consolidation feature is probably the thing that stands out as the most compelling because that's really difficult to do. And there's very few that have done that very well in terms of building on existing integrations across the major accounting platforms.

[00:49:51] And now they can project group figures and handle things like eliminations, manage the parent subsidiary structures within the same environment. And having been a CEO of a multi-entity and with multi-jurisdiction, it is very difficult to then manage that across what could be seasonality and cash flow and how that translates to how the workers in certain countries work.

[00:50:18] So I think this is crucial incoming financial data that will be modeled in a much more sophisticated way. So the main takeaways are pretty straightforward. They bring the integrated three-way modeling right alongside the real-time reporting and cut out the need for constantly rebuilding custom spreadsheets. What do you think, John? I was going to call you Don then. I don't know why.

[00:50:45] What do you think Don doesn't approach like this genuinely unlocks the bandwidth of the practice? Or does it just simply reallocate human time towards validating parameters? So I'll avoid the obvious joke, right? Since I smirked whilst you were saying it. But so I think some of what's just been talked about there is just what I would expect from a forecasting tool, right?

[00:51:10] You know, if I'm changing a revenue model in the P&L, I expect the cash flow statement to reflect those changes in the balance sheet as well. So a proper three-way forecasting model should be fully integrated. And that is one of the challenges of doing this in Excel, unless you build a really clever model. You can have that disconnect that they're talking about. But I think when it comes to software, my expectation is it should do that.

[00:51:34] So what really appeals to me in this announcement, though, is there is an element of agentic stuff in there. So you can actually chat with the model and you can sort of tweak things as you go along. And actually, I do think that is quite a handy interface sometimes, you know, in terms of just being able to say, hey, can we try this with a change in seasonality or a change in scope, you know, and see what the outcome is and do that. Rather than having to model it through numbers and shifting things around on a screen and stuff,

[00:52:02] I actually do think that's where a chat interface can help quite neatly. And it also maybe gives you the ability to kind of like rip a model that you've already got and like just play around with it in a sort of a separate environment without destroying what you've already got. So that appeals. And then the group stuff that you talked about, massive, massive for me, because, you know, even if you're using a piece of software like Join, historically, you would have been doing that one by one. So you've got one entity here in the model. Then you've got another entity here in a separate subscription or a separate directory or whatever.

[00:52:31] And then trying to aggregate that together is probably you just chuck it all back into Excel and mash it together and hope it all works. So being able to do a kind of consolidated forecast across a multi-entity organization, for me, is like a massive, massive win if it's reliable and you can get all that output back in. You know, and so that definitely, definitely appeals. Do you think it's still underserved in general from what you've seen in the market?

[00:52:59] I think multi-entity broadly is underserved. Yeah. I mean, obviously, there are some really good tools out there and we talk a lot about them on and off over the episodes and stuff. But I do think, you know, I guess you've got two things with multi-entity, right? You could be multi-entity all on the same platforms. So you could be multi-entity on zero, for example. Or you could be multi-entity on multiple platforms. And that has its different levels of complexity and challenge. I also think you, despite all of the great tools that we've talked about, you know,

[00:53:28] whether it's around reporting or forecasting or broadly, like helping to manage that multi-entity stuff, I still don't think that anyone's really, really nailed it down. And like, particularly, like, you know, even on like Mayday, like, you know, when I use Mayday, I like the product. But I still look at it and think this could be better. You know, I think there's still like missing bits, missing middle, there's missing, you know, missing tools or the ability to do certain things that would make my life a little bit simpler. And so, yeah, there's definitely some stuff there. And the big challenge for all of these tools is integration, right?

[00:53:58] Because once you go like multi-jurisdiction, multi-entity with multiple platforms, you know, like I know when I used to work in audit across sort of, I can't remember how many it was, like five, 600 clients, we had like something like 130, 140 different accounting systems in use. Just imagine how bonkers that is trying to like, you know, integrate or pull the data from those to make it really work. We live in unprecedented times. There you have it.

[00:54:24] There's another episode of the DigiTools and Accrual World podcast in the can brought to you, of course, by The Loop. And as Indy said in the intro, it's so good actually for all three of us to get back together. It doesn't happen very often. We're usually just like messaging each other on WhatsApp nowadays. We barely talk because we're all so busy, but it's really nice to get back together. It's disappointing to miss some of our guest presenters this time around, but I will always the next week. And yeah, I think it was so interesting. Lots of stuff to talk about. I think, like I said at the beginning, I feel like the AI debate is starting to mature a bit more.

[00:54:52] I think we're having better conversations. And there's just so much at stake now in terms of like where the technology platforms are going and like where broadly like the thinking about what the accounting firm of the future look like as well. And like, how do we model that? I model that with all the other challenges that we have around the periphery, you know, because none of those other challenges have gone away. So yeah, fascinating times. Of course, as ever with the podcast, we'd really like your feedback, your thoughts, your ideas, what's good, what's bad.

[00:55:21] If we've missed something that you want us to cover, of course, let us know as well. You'll find us on all the usual channels, LinkedIn, YouTube, all of the podcast places. And yeah, we'll see you on the next one, right? Next one.