Accountants (Apparently) Optimistic About The Future || Accounting Tech News

[00:00:00] They are no longer just a zero product, they are integrating into QuickBooks. Nearly 80% of firms reported an increase in revenue and nearly 75% reported an increase in profits. The question that we see from most people is, right, okay, it's great to talk about the technology, but how do I utilise it? How do I implement it? Three quarters of the firms are massively optimistic about the future and 25% are just there going, you know, we'll see how it goes.

[00:00:24] AI is still an infant. It's now the time to bring it up with the right morals, with the right ethics, and we're not doing that. I was working with AI before AI was cool. Now it's very cool. Everyone that I say I'm the AI leader, they were like, oh wow. Back in 2020, I had to explain to people what I do. It's been a long ride and it's only with the rise of large language models that it became so viral, so famous.

[00:00:48] And we could not even predict it, to be honest with you. The most important skill that we need to have right now is... Hello and welcome to the Digital Always in a Cruel World podcast, the place to go for your insights, news and information on the digital world of accounting in the UK and beyond. This episode is brought to you by AdSum. If you've not caught the recent Digital Digest episode, then you want to dive into that to find out how AdSum have released corporation tax into their open tax product and how that's helping accountants across the UK.

[00:01:16] Okay. We'll also be diving into an interview on AI with Monica Odysseus. Check that out with John too later. And as ever, App News bring you all the insights to what's changed recently in the UK accounting news. But before we dive into all of the detail, let's hand over to John and find out how he is, because it's just us two again today.

[00:01:36] And we are amiss. We're amiss without Indie Tatler. After the stonking hangover she had for the last episode, we've now had account tax and I think she's still recovering from that. Yeah, I think she must be. I mean, well, you know what Indie's like, she likes to go hard. So, you know, I'm sure she'll have got the most out of account tax that she possibly could.

[00:01:57] But as always, you know, great, yeah, great event. Lots of people, lots of interesting talks, lots of, you know, fascinating insights to be found from some of the vendors and some of the speakers as well. And tons of tons of things to do. That's always a challenge with account tax, isn't it? You know, it's a lot of people there. It can be quite overwhelming.

[00:02:20] It's a fairly full on couple of days. Hopefully we helped a few people on the automation station with a few ideas and thoughts and you're pointing them in the right direction. At the very least, that was that was the goal. And yeah, I think we'll all be needing a little lie down for a few days just to recover, I guess. I do. Being as old as I am nowadays, my back is particularly painful.

[00:02:44] But I think the other thing I loved about account tax was the launch of our kind of new training platform or training scheme, which is helping those who want to launch or run a digital transformation service line in the accountancy firm. We touched on that. We gave people insights into an element of what will be rolling out over the coming months. So if that is something that you're trying to do, want to do, want to grow, please do reach out to us.

[00:03:10] But we always find when we've been at a conference, we've talked to lots of people, we get a spike of new listeners. And so if you are a new listener, thank you for reaching out, listening to our podcast. We love to hear from you. So please do. At the end of this, let us know what you think. And if you want to be part of the Digital Disruptive community, drop us a message. We love to hear from you. Best way to reach out to us is on LinkedIn. But let's just jump straight into it, John. As we always do, let's jump into talking about tech and find out what's been changing over the last week or so.

[00:03:43] Amazing. So I'm going to start off first with a little bit of research that I picked up from a group that I haven't come across before. But, you know, that doesn't mean anything really, because it doesn't mean I've got my finger on the pulse when it comes to research. But there's a group called Credence Research who released a report on the accounting software market, particularly on the small business side of things. So looking at sort of your typical things, Xero, QuickBooks, Free Agent, Soho, you know, et cetera, et cetera.

[00:04:13] All the usual names that we're kind of familiar with. And I guess what I'm going to do, because it's quite a long thing, and they've got a very detailed research report that you can buy for excessive amounts of money, because it's based in the US. But I'll summarize the main findings that they published, which I thought was really, really interesting. So they reckon that in 2024, the market valuation for the small business accounting sector was just shy of $24.5 billion US,

[00:04:40] and that their projected value for 2032, so what's that window? Eight years, or seven or eight years, is to climb to just shy of $51 billion. So just over double in size, so growing by about 10% in that time. So yeah, that's a really interesting, I guess, thing to see. And I guess very exciting if you're a Xero or Intubit shareholder, for example, because you're going to see more success and more profits hopefully coming down the line.

[00:05:11] And they're sort of saying there's a lot of reasons for this growth, but in particular, it's probably a lot of surprises, but a lot of that growth has been driven by the adoption of automation. And that obviously has an impact in terms of reducing the errors and enhancing productivity. And in particular, businesses are really focused in on automating sort of like the typical bookkeeping tasks, I guess, that we're used to, and then some of the follow-on tasks. So invoicing, payroll, but also then tax calculations and reporting out the other side.

[00:05:40] So perhaps not surprising, North America is the largest market by a fairly long way. It has about 35% of the overall market. I guess that's probably dominated by Intuit and their position there. What I thought was really interesting was that Asia Pacific comes up next at 25%. And I suspect, although there's not enough analysis detail in there to kind of get into the ribs of that, but I suspect that's possibly driven by Xero's presence in Australia and New Zealand, obviously.

[00:06:07] And then they talk about the impact of larger countries like China and India, and obviously a whole bunch of other big countries that fall into that segment, that are starting to grow and expand their horizons and looking at software solutions to help with that. And clearly, there's a lot of population there in terms of headcount to change and improve the way that they go. And then Europe came in at around about 20% and then Latin America after that.

[00:06:36] And there was a real sort of interesting review of the fact that the European market had been a little bit slow to start, but was starting to pick up more and more. And then the key market drivers, which we talked about very briefly, is that, you know, not too surprisingly, people are wanting to move to the cloud. We've talked about that lots and lots of times on the pod in terms of, you know, how slow that can be sometimes and how we want to encourage people to do that more quickly. There's, you know, a huge opportunity around automation.

[00:07:05] And obviously the AI conversation bleeds into that in terms of being able to do more and more with that automation opportunity. But that also improves the way that you can look at your data and maybe improve the way that you go about making decisions and stuff. And then the final bit, which I think is probably particularly prescient in the UK market, given the whole MTD agenda, is that, you know, there's increasing levels of regulatory compliance requirements,

[00:07:29] which are encouraging businesses to move to utilizing accounting software to make sure that they stay compliant. And I guess you could kind of overlay that with the, you know, e-invoicing requirements that we're seeing in lots of other industries as well. So it's not just tax, but typically it is. And then I guess after that, Ian, boy. Wow. So all good for those that are investing in accounting tech, basically. Yeah, absolutely. But I guess, you know, you flip down its head and sort of say, well, what is the opportunity for accounting firms?

[00:07:58] Well, I guess the opportunity for us is, you know, around advising our clients, helping them to navigate this space. I mean, it's not a new opportunity necessarily, but if they're expecting this market to increase by 10% every year, then our clients are going to be fueling that. And so isn't it better for us to serve our clients rather than maybe, you know, an IT business or someone else serving our clients and guiding them through this decision-making process? Yeah, I'd say 10% feels small.

[00:08:26] Not for, I guess, the wider remit, but if you're a zero, you'd be thinking, you know where we've been? 10% is small. We want a bigger chunk than 10% pushing that. But you talked about the impact for accountants, and Xero did release their accounting and bookkeeping industry report not that long ago. And I wanted to dive into some of the stats that came out of that and the impact that that may have on how we project into the future and what we should be looking at. So if you've not come across one of these reports before, I know we talk about it possibly annually,

[00:08:55] but at least a few times on the pod. This is an online survey, and the 2025 report covered around 250 accounts and bookkeepers from, I guess, independent panels. So select people, and they're trying to do a spread across large, medium, small, and sole practitioner businesses to see what the impact has been. And the biggest impact we've seen is it's been a very good year up to 2025.

[00:09:21] 80% or nearly 80% of firms reported an increase in revenue, and 75% or nearly 75% reported an increase in profits. And that's quite a big chunk. We've been through a lot of turmoil. We had those numbers closer to 50% of times. So the fact that now we're up at 75% to 80% for improving the money that we're making as firms is a big shift. And I think that that, I guess, reflects the changes that have been going on.

[00:09:51] There's been a lot of flux regarding acquisitions. But also, one of the biggest drivers that we're seeing for firms making increased revenue is from the acquisition of new clients. And nearly 50% of firms increased their client base in that year, with only 8% decreasing. Now, obviously, generally, no one wants to be in the decreasing side, unless you're scaling down. But the fact that there are constant new businesses coming in,

[00:10:19] and we're constantly growing as an industry in the UK market, is very impressive. And then they touch on certain things that they always like to get out there. So if you are a cloud-first firm, then you are seeing a much higher level of growth, 51% compared to 38%. And if you are a zero-powered firm, then you're getting 57%. You're more likely to be growing compared to 42% that don't utilize zero. So, you know, big shifts there.

[00:10:48] And that's just those that use zero. If you are a fully integrated zero platform, then it's 70% of those showing the biggest growth. Now, you know, you always do wonder, is this a one-sided survey? It's hard to tell because it's run by zero. So it's going to be putting those positive figures out there. But we do see a trend of that. Whether it's zero or Intuit, if you are a firm that, I guess, is thoroughly behind the software vendors

[00:11:14] and using that as a big area for pushing your business forward, then generally you're going to see a big improvement. And then some of the stats that go behind this is that, well, what's driving revenue? What's driving profit? And the biggest gain for revenue, I guess not really that much of a shock, is increased number of clients. Well, kind of obvious. And then growing relationships with existing clients. Well, obviously quite obvious as well. So there's not really anything that comes out of that.

[00:11:44] But from a profit perspective, the number one change has been an increase in the number of services provided. So growing what we deliver rather than doubling down on what we're already doing. And so I think firms are now trying to expand, do more rather than, I guess, niching. And maybe that is then helping with that profit growth. And then the number one change or shift for the ability to take on new clients. So as I said earlier,

[00:12:13] you're seeing like 50% of firms that are taking on new clients. Obviously, the availability clients is the driver of that. But you need to have the capacity to do it. And the biggest driver of the ability of that is increased efficiencies, allowing firms to serve more for the same pool of individuals in their firm, which we are seeing. It's a general trend, John, that we've been talking about for quite a while, is that that balance in costs between employees, human cost and tech cost is shifting.

[00:12:41] You are reducing your employee cost and increasing your tech cost, but still are the same way of improving your profit. So yeah, some interesting stats out there. Would I say there are any massive shocks? No, but no, I think it's very clear now, and it has been for a while, that if you are cloud enabled, you're cloud powered, you're going to be doing a bigger growth, better profitability than your competitors. At every event and every webinar talking about AI,

[00:13:10] the question that we see from most people is, okay, it's great to talk about the technology, but how do I utilize it? How do I implement it? And I think this is one of the challenges that we're constantly seeing. It's not new. We talked about this a few times, but when products like Xero came along and QuickBooks launched the online version of their product, was that we knew what that was straight away. It's a bookkeeping and accounting system. We know how to use it. Yes, the buttons are different colors and different shapes and in different places, but it made a lot of sense to us.

[00:13:39] And the challenge around the use of, well, if we're talking about AI, the conversation's always large language models. And the challenge around using large language models is that it's not a product that you just utilize off the shelf that solves a particular problem. You know, it's not like the apps in the Xero ecosystem where they're coming in and saying, right, we can help with your approvals or we can help with payments or we can help with processing documents. This is something where as accountants and businesses in general, having to go, right, okay, well, this I can see is a great opportunity,

[00:14:08] but actually let me now go and find a problem to solve with it. And how do I make sure that I do that effectively and efficiently? And do I have the time and the capacity and the capability and all the other things that go with it to be able to do that, you know, in an effective way? So yeah, not too surprising, but we'll see more of more of this and I'm sure we'll see more use cases as this level of maturity of utilizing large language models continues. Definitely. And I guess the last bit I just wanted to pull from this report

[00:14:36] is something I'm surprised at from the questions I've had and the conversations I've had with other accountants is that 75% of those that responded were positive and optimistic about making tax digital for income tax. And that does seem to go in the face of a lot of the other things we hear out there is that people are just, I guess, a bit frustrated with it, you know, just dragged on and then they're trying to get their heads around how they service their clients effectively. But yeah, it's 75%.

[00:15:06] I mean, that's quite a high number. And most of that seems to be from the potential to win new work. But who from? Like if everyone's doing this, who are you winning the work from? Is it that we think every firm is so optimistic about how they're going to deliver it? They're going to steal clients from elsewhere. It's an interesting thought process around this. And maybe it's, we've convinced ourselves because it's been such a frustrating period that actually this will be good. We'll make this good because we are optimistic. We're, you said like, well, the only, one of the other stats in here is that we were optimistic about the future.

[00:15:36] You know, there was a, you know, 77%. And you can see these stats are always generally around 75% to 80% that are positive. And maybe it's all the same firms that three quarters of the firms that were surveyed are massively optimistic about the future. And 25% are just there going, yeah, you know, we'll see how it goes. Moving on from the sublime, maybe to the ridiculous, if we're talking about mindset and technology, HMRC are looking for a new CRM, Ryan. I mean, this sounds like

[00:16:06] a cracking opportunity for you and I, maybe, to be involved in some transformative change at the tax regulator here in the UK. But they have, and I mean, you know, look for a massive payday with this, by the way, Ryan. They have earmarked half a million, half a billion pounds, sorry. Yeah, I think half a million doesn't feel like much. Half a million, yeah, sorry. I forget you don't get out of bed for less than a million. So half a billion pounds they've earmarked

[00:16:34] to move across to a new core CRM SaaS platform and they want that to be able to manage the registration, subscription, and customer record management capabilities, effectively replacing a whole bunch of old software. And what's even better is they've earmarked another half a billion pounds to pay someone to help them with this. So I mean, this is just absolutely staggering really. So this has just been launched out in the market. They've sent out what is called a preliminary market engagement notice

[00:17:04] to effectively that allows them to scope suppliers to provide the software. So I'm not sure who will be involved in that, but I think you can imagine all of the major like big CRMs and probably some of the big ERPs may well be involved in this. As always, we always know that the outcome of those implementations always ends up being terrible. So I'm sure when the Public Accounts Committee come to review this in a few years' time, we'll find that HMRC spanked a load of money on something that didn't really work

[00:17:34] and that hasn't been maybe as successful as the outcomes. So that's probably the key thing is they've got to be able to manage this in a way that's effective because, you know, they, you know, in our experience and the experience, you know, look at Birmingham City Council and all those others is like, be very, very clear about what your requirements are. Be very, very clear about the fact that you should be utilizing the core functionality within a product and don't, don't try and bastardize the product or don't try and tailor or tweak the product too much to your capabilities

[00:18:03] because if you're doing that, you're just going to be locked into a world of pain and you won't get the outcomes that you want. So that's a key thing. I mean, but one of the other really, really fascinating things about this is because the Public Accounts Committee had done some research which is included in this and sort of prompted one of the changes is that, you know, they talked about the HMR phone service which I think quite a lot of accountants of bookkeepers have been the victim of over the last, over the last few years. But in the first 11 months of 2024,

[00:18:33] they cut off nearly 44,000 customers who've been waiting for 70 minutes to speak to an advisor because the system couldn't cope with demand. Only two thirds of calls were answered and the average wait time was more than 23 minutes which is just, you know, absolutely, you know, bonkers. And if you want an even more staggering fact or statistic from this was that if you were phoning in for an inquiry with HMRC in the fiscal year 2023,

[00:19:03] you were on hold for a total of 798 years. I mean, that is absolutely mind-blowing. And in that fiscal period to March, so that's the year ended 2023, March 2023 is where these numbers come from. there was more than double the time wasted than in the equivalent year to 2020 which is, you know, just shows that like HMRC

[00:19:33] have just completely lost the plot over the last few years and, you know, investment in technology but also in people and skills because there's been a whole brain drain from HMRC is desperately, desperately required. Yeah, I mean, they need a new CRM even if it's just so that you can link one, you know, your business in one part of HMRC to the other in an effective way. That as a minimum would be quite an achievement. Yeah. And yeah, I mean,

[00:20:01] it's half a billion, right? That's insane. Although, for context, John, what they're saying is the impact of making tax digital on businesses is half a billion for income tax. That's the increase in compliance across all the businesses that are going to be using this is half a billion. Wow. If only they could use that for some sort of, you know, system that they can impose to make things better. But unfortunately, this doesn't go to the government.

[00:20:31] This goes to the consultants, the advisors, the accountants that will be helping those businesses. Hence, probably, why 75% of firms are optimistic about making tax digital for income tax. But they've already said that from the making tax digital for VAT, because obviously they're predicting these figures, but for making digital for VAT, the impact's been for 300 million. And those are businesses that were already mostly quite tech savvy, had a solution that really, just by turning on the filing system, enabled them

[00:21:00] to file their VAT records. And that impact was 300 million. For self-assessment, the fact that most of these businesses or a large chunk of these businesses wouldn't have had software to help manage in their finances. Yeah, not only the impact or a cost perspective, but the impact in disrupting exactly how they operate, their relationships with their advisors is huge. And I know that they said right at the start when they released this, that this would help overrule and that they expected that the savings, efficiency savings to those businesses would override

[00:21:29] the increasing cost. And this does not reflect any, I guess, efficiency benefits by businesses adopting effective technology, nor the improvement to make a better profit by the increased visibility of their financial reporting. It's hard to put a figure on that, but I'd be surprised if it's half a billion across all taxpayers. So, I think it's going to come across as a net cost at least

[00:22:01] if HMRC can get a new CRM in, they can get some visibility and make some better decisions based on the information they're going to have from these schemes. That could be a positive, I guess. Absolutely. I mean, yeah, I mean, it's perhaps not surprising that there is a compliance cost as a consequence of various strands of MTD. I guess what we don't necessarily have in the numbers or at least I can't do the maths in my head very quickly is what's the individual impact per business? Because, obviously, it sounds a lot half a billion

[00:22:30] impact for all of these organizations, but you've got to remember there's quite a considerable number of organizations going to be impacted. And so, when you kind of divide that up, you're going to have a relatively small impact per organization. And obviously, that'll be different based on your type and your size as well, of course. Yeah, perhaps inevitable there's going to be a cost to comply and that cost falls on the businesses and comes back to us probably as advisors, which you can't necessarily complain because that's

[00:22:59] economic expansion and growth opportunities if you want to dress it up as a positive. But, yeah, there you go. Moving on, I've got a fundraise announcement from Canopy, which is a US-based practice management solution, perhaps not that well-known in the UK. I don't think they've got too much market presence here, but I've bumped into the team a couple of times on my travels and stuff. But they have raised $70 million

[00:23:28] in a Series C and guess what, Ryan? It's going to be to help them expand their AI-driven capabilities. What? No. I know. Who'd have thunk it? But, yeah, I mean, look, let's not be too silly here. Let's congratulate the team. I mean, $70 million Series C is pretty impressive. It's been led by Viking Global Investors and then some of their existing investors, Tencoves Capital, Encona Capital, Pelleon Venture Partners and Tenaya Capital.

[00:23:57] This is like a word soup for, this is a challenge for you, Ryan. We're also part of the fundraise and this is double what they raised back in May 2024. So a year on where they raised $35 million and they've raised another $70 million and their total fundraising to date is roughly $285 million which is great news for them and clearly they're doing good things to make that, to make the business

[00:24:27] investable and to continue growing as well. So there's a few bits and pieces from the CEO from Davis Bell who sort of said they've been talking to Viking for about a year to encourage them to invest and want to demonstrate their success and things. So this is just the world that you've got to be in I guess when you're building software and obviously talking to potential investors is part and parcel of what you do and I think the key thing for them is like one of the really interesting things I think they want to focus

[00:24:57] in on is around how do you go about that initial engagement with your clients when you're requesting information and sending them information as well how do you make that as seamless as possible because I think that's one of the areas where AI could definitely help but I think it's also an area which no one has managed to fix yet in terms of that correspondence between ourselves and a client particularly when we're requesting information because we know that sometimes clients either can't find the information or they can be a bit unreliable and you spend a lot of time going back and forth going right

[00:25:26] okay I've asked for this you sent me this as a response but it's not quite what I wanted so I'm going to have to go back to you and then refine my request and tweak it a little bit so I think there's some great opportunities there and yeah the key thing that Canopy do let me just summarize this for you is they serve currently about four and a half thousand accounting firms they are SOC2 compliant so that means that their data is super secure and the platform is super secure they particularly cover the whole client

[00:25:56] engagement piece document management workflow time billing engagements proposals and more so you know I think these guys should be looking to the UK because there's plenty of competition when it comes to practice management solutions here but also plenty of opportunity because I don't think anyone's quite nailed it yet no I was thinking just that John is that I bet we'll be seeing them over in the UK in the not too distant future but another bit of information on news that isn't UK focused at the moment

[00:26:25] but will be in due course is that Xero have partnered with BGL work papers now you may not have heard of BGL they're not a business that's well known in the UK but are huge over in Australia and what Xero effectively are doing with this is they're identifying that they need to reimagine how they do work papers and rather than do that themselves they're going to work in partnership with a leading work papers provider over in the Australia marketplace

[00:26:55] now they focus on BGL being a leading compliance management and AI powered paper to data software provider but essentially you're looking at a business that has built really good systems such as cash 360 and simple fund 360 to help those over there and to build out their accounting compliance and has already come over to the UK with cash 360 very recently so what do we expect from this well obviously they're going

[00:27:25] to focus on a single source of truth that's something that is very critical in the way that Xero work so going in from a working trial bank all the way through into your working papers and then through to tax returns and financial statements something that has always been a bit bumpy I think with the current Xero work paper solution so they talk also about the deeper integration into reporting and zero tax as well as managing the practice client records and the whole security regarding staff permissions

[00:27:55] but then as well as that enhancing collaboration so if you ever work with work papers there is an approval and review process in there but it's probably a little outdated now compared to other tools that are in the market that make it a bit more seamless so how would we expect this to roll out so there's a beta that's going across in the Australian market early on so you can elect to be involved and then that will be followed up with releases for practices in New Zealand

[00:28:24] and the UK so they do make a specific reference to us seeing this over in the UK and I think it's going to be important for us John to keep our ear or nose to the ground with what's going on over in the Australian market because we think that this may have a big impact over in the UK because Xero have never really pushed the whole practice management work paper side that heavily we're starting to see it a lot more with Xero tax I think they've been a bit coy about Xero work papers because of what they saw and felt about the product and what

[00:28:54] feedback they probably had from the marketplace but I'd expect this to become a lot more aggressive as a promotional side once the BGL powered work papers comes over to the UK it's going to be interesting to watch that for sure absolutely well since we talked about Xero might as well talk about one of their big competitors and QuickBooks have announced a whole bunch of updates and improvements to their sole trade product and the key one I think

[00:29:24] which for me stands out at the very least is that you've now got account and access in the product so it allows you to have a bit of a summary overview of your clients that are using QuickBooks sole trader so that you can navigate between them and switch between clients a little bit more quickly than you could do beforehand they've then added some additional bank reconciliation functionality so you can add statement balances match transactions and then also help to see

[00:29:54] if anything has been excluded just so that you can review that and just make sure it's correct so that's really important because as we know lots of sole trader clients tend to mix their business and personal banking together and obviously you don't want to be including personal transactions in the MTD submissions that you would be doing in the platform going forward and then now you've got some additional functionality around chart of accounts and mapping so you can add additional nominals to the chart of accounts which you

[00:30:24] weren't able to do before and then you can map that through so that you know is this a new income code or a new expense code for example and where that would go in the detail and on top of that they've now announced the ability to do some additional reporting and create work papers so some basic reports which weren't there balance sheet and P&L allow you to sort of see how that sole trade business might be performing and you've also got the ability to utilize integrated work papers

[00:30:53] which you'll be familiar with if you're using QuickBooks already because it's kind of there already in the side products that sit alongside this but you've got now the ability to review work papers document sorry attach documents and see a bit of an audit trail of who's prepared what and then finally some additional things if you're in practice it just makes it easier to collaborate with your clients because you can request information and add notes for queries or anything like that and you can also close out work papers

[00:31:23] production and completion utilizing some of the functionality I mentioned before so again just helps you to keep on top of a whole bunch of things so this is a welcome additional functionality to QuickBooks trader why it wasn't there maybe in the first instance is a question to be asked but look it's all part of the process I guess as people are getting more and more used to these kind of platforms building on the MTD requirements that firms and their clients are having to start dealing with particularly

[00:31:53] as we're now in this new tax year with the new beta has opened up all very interesting and we talked a lot about the main providers but let's close this episode out from an app news perspective today with an X system one that used to be owned by Zero but is now owned by BlueRock which is Workflow Max they've been constantly releasing new features for their product and there's been a whole host that have come out recently I think mostly through March it's a little bit outdated now

[00:32:22] but all still quite pertinent so they released multi-factor authentication and mandated it and that should now be in the product so if you are utilizing work and actually be aware of this but if not then please make sure that you are happy with a mandated multi-factor authentication across your system generally this is I think probably something that's kind of the norm now it's something we probably expect in most software and something that John doesn't stop ranting on anyway

[00:32:53] so it's good that they have brought that out to flag it as to not be invoiced and similarly you can add something to a job where you want it to be included in invoice but not to go into cost I can't think of a scenario that you'd want that at this point but I guess you don't want to track it from a profitability perspective then that is something you've got

[00:33:22] functionality to do on a cost by cost perspective you can bring in notifications for any lead tracking which I guess is something I probably would have expected to be there in the past but it must be something that was missed because this is something that you don't want to lose sight of you need people to be notified when there's new leads and then you've got improved flows for the importing of Xero contacts for handling job management and job financials and then I know that they're bringing out a deeper

[00:33:52] integration with Xero over the coming months as well beginning from April but going through the first quarter to improve how the invoicing approval process works and greater transparency on what has synced and what hasn't synced so lots work work there but the biggest one is going to be that they are no longer just a zero product they are integrating into QuickBooks so you're going to have a QuickBooks integration with Workflow Max that will work exactly the same way that Xero works

[00:34:21] so focusing on that task job management and building out on the QuickBooks platform I was working with AI before AI was cool now it's very cool everyone that I say I'm the AI leader they were like oh wow but in 2020 I had to explain to people what I do okay cool I am joined today by Monica Odysseus who I know through our connections at ICAW we both sit on the data analytics community advisory board which is always exciting

[00:34:50] I mean my involvement is way less interesting than Monica's because she's a guru and one of my most favorite people that we get to involve on the committee and stuff so it's really cool and I should just make sure I had to write this down because Monica's title is like super exciting and it's probably more meaningful than mine because mine's a little bit vague but she is the AI and data lab leader at Grant Thornton in Cyprus so Monica delighted to have you with us maybe you could give us a little bit about your

[00:35:19] background and how you've kind of ended up being in this like AI leadership role perfect well thanks for having me for starters and I'll give you a bit of a background because it's quite diverse how I ended up here and quite exciting of a journey so I was always since a child I was always interested in many different things so I was looking constantly to learn something else so that gave me the diverse mindset let's say

[00:35:49] to explore different stuff so I started my career with a bachelor's degree in mathematics and computer science I started my first job after that was as a software developer which was my favorite subject at university because of the problem solving mindset that it wanted and that I loved and then following two years after that I did my master's degree on risk management to combine it a bit with business and then started

[00:36:19] got a job at PWC where I started my qualification my ACA with ICAW to become a chartered accountant and having no idea what a chartered accountant was until that time I had no idea what debit and credit was so it was all new to me it was all very exciting it was I was a bit angry at first not understanding where debit goes and where credit goes but got the hang of it with the passing of time

[00:36:49] and it gave me a really nice combination of business and technology that I had so I really like that combination so I was looking something to combine the two and it was in 2020 that following many positions that I had at PwC moving through various departments it was in 2020 that a new department called the AI team opened up in PwC and that's when I decided I wanted to join and which gave me

[00:37:19] that combination of advising so business understanding business which I wanted to use and the technology concept as well so I was working there for four years and it was one of the best jobs actually I had because it combined gave me the problem solving stuff but also the business mindset and it's an industry where it's constantly developing

[00:37:48] it's never the same like things I knew and I was working with a year ago do not exist now so we are constantly reinventing ourselves so last year an opportunity came from Grand Thornton Cyprus to open up a new team at Grand Thornton the AI and data lab and I was called to initiate this new department to lead this department and created a team from scratch and we're now one

[00:38:18] year in and it's very interesting very exciting and we're very hopeful for the future as well and maybe just before we move on I mean just give us a little bit more detail about what the data lab is there for I mean why are GT as an organization you know clearly they're going to be interested in AI right but why are they interested in sort of setting this up as a separate team a separate department and what are the objectives that you're trying to achieve over the next few years with that okay so as a team we're working we're under advisory

[00:38:48] so we are advising clients on how to use AI within their organization which are the best solutions to apply how to utilize their data to get some insights to get value and improve their decision making become more efficient and depending on the needs of each organizations we apply AI or data analytics data science depending on each condition we're also involved a lot in upskilling because I really love teaching not teaching but I would

[00:39:17] say training people I really love making complicated topics simple that's my trademark let's say I like making things simple that's why I wrote many articles on how to what is AI really for example so we are a lot involved in upskilling people and organizations as well through trainings and yeah the reason why Grant Thornton wanted to open such a team very good question it's a very

[00:39:47] forward looking organization and very focused on finding niches on the market and investing in that so they have other teams that they invested in as well for example quantitative risk and cyber security which are their niche services or ESG and they invested because they see the value and the potential in investing in such solution and this is what I really liked about

[00:40:17] Grant Thornton's philosophy let's say the investing in startup ideas and investing in niche things with a forward looking and a vision to the future amazing amazing well you're in the perfect place to keep things simple with me Monica so that's all good so you said you got into this in 2020 with PwC five years ago now and I guess we constantly have this

[00:40:47] interesting battle when we talk about AI because whenever we go to conferences and events now the main conversation about AI is pretty much about large language models which is really just one small branch of AI isn't it the point I'm trying to make here is that five years ago large language models were a little bit of a pipe dream and certainly weren't the commercial success that they are now but AI back then in terms of a conversation probably wasn't really that much on the agenda even though most businesses and most accountants were

[00:41:17] probably using elements of AI maybe without realizing it and I just wonder do you think move the AI conversation front and center in terms of the impact definitely well I was working with AI before AI was cool now it's very cool everyone that I say I'm the AI leader they were like oh wow back in 2020 I had to explain to people what I do so

[00:41:47] and we were working with there's a misconception of what AI is we have many clients coming to us and saying I want to apply AI in my organization and what they really mean is I right and it's only with the rise of

[00:42:17] large language models that it became so viral so famous and we could not even predict it to be honest with dishes out of the dishwasher or folding

[00:42:47] our clothes would come a lot faster than a model who will be able to communicate because it's such a complicated thing to develop we see now that it was life changing

[00:43:16] for people and for organizations we are reinventing the way we work but over the next years I mean now it's the rise of the AI agents who are going to get a task and then go and follow whatever they think it's true to follow in order to achieve that goal that is given to it we are going to see a lot of improvement and a lot of reinforcement learning as well in the next years so and we cannot even predict I don't believe anyone can predict where things

[00:43:46] will go so it's only guesses that we can do yeah I agree and I think I so difficult to actually nail down a prediction and be relatively reassured or accurate that you'll get that right one of the things that really impressed me when we I think we were together face

[00:44:16] to face which was last year at one of the ICAW conferences you did a do some of that predictive work that allows them to have these natural language conversations and stuff and I wonder could you maybe just touch on that a little bit and give us some analogies about

[00:44:46] how do these things work and then maybe we can understand them a little bit better question okay let me think how to make this simple remember who you're talking to so the way that these models work it's the data that we have provided to them whether that is data that we have online such as Wikipedia books etc the web pages blogs etc and

[00:45:16] the data that we share on social media that it was able to gather all this data large language models were able to gather all this data and use them to understand how we structure our sentences how we how a context is defined and so that was the data set let's say so using that data set models were trained to understand how these things work

[00:45:45] using supervised learning and then following that some other techniques of artificial intelligence like supervised learning reinforcement learning was also applied to check which statements to rate which statement of the models which one performs best in order to assess how to make this model better and better and then it was the first I think in 2018 2019 no when was it GPT

[00:46:14] 2.5 I think was the first one that got out it was only for researchers and then after that they were building on it when they

[00:46:44] became larger and they were an interesting fun fact if I may share was when they asked the CEO of Google at the time Larry Page they asked him why do you offer Google search for free and what is in it for you and he said that Google search is not really what we are trying to aim this is not our purpose our purpose is to build an AI so data is the

[00:47:15] largest asset that we have in this era and without data we could not build large language models and I think that's why we saw it so much advancing in the last years it's because this social media explosion the connectivity between us and also the improvement the advancement

[00:47:45] of computing power which is increasing by 30% each year so advanced computing power gave the right foundation for LLM models to be developed yeah and can you sort of explain what is a hallucination I mean

[00:48:14] hopefully that's self explanatory but why do they come about what is it particularly about these models that does create these hallucinations as a consequence of the way they work well the models as I said they are trained on the data they are given so if it's not like the models are bad in returning stuff most probably the if the data let's say that we gave chat GPT or

[00:48:44] Gemini or an LLM to train on were biased then the model will be biased it's not the model per se that makes the mistakes but the data was fed on and the process so what AI hallucinations are it's when you ask something a large language model and it replies back with such a convincing way that you believe whatever it tells you and even us that we are working with AI all

[00:49:14] day and we know the risks and we know what might go wrong and we know what AI hallucinations are we find ourselves I found myself many times being convinced by something that going on and like redistributing that information to people and then when someone came back and said you know Monica are you sure this is not it is right and then I had to go and cross check and it was not but it's so convincing the way it does it even gives you

[00:49:44] sources that I found this yes I'm right I given and not to I always talk about large language models and we use chat GPT in our firm at the moment but I

[00:50:14] always talk about it like the overconfident trainee right they critical thinking comes with experience let me give you an example we were looking at VAT registration and stuff here in the UK and some

[00:50:43] people got responses that had old information in there and other people got a different response but no one really did that step back is that right am I absolutely comfortable that information is reliable and what I'm getting is right so do you think there is an element of this comes from your professional experience as well definitely knowing how to research also really helps so

[00:51:13] knowing how to cross check information what I really found that it's useful it's you don't trust the context you trust the source of the context so if chat GPT tells me that I found this information and it's true because I found it from the onion news dot com dot whatever then I would be like the onion news really okay whereas if it tells me you know what I found this

[00:51:59] information that you know that that's become a huge topic of the last few months or even probably more so but as you

[00:53:00] people to know what can't protect but because things are moving so fast and regulation is moving a lot slower we don't have an answer to that and we don't have an answer to we cannot protect writers we cannot protect

[00:53:28] content creators we cannot protect artists because all this information is used by llms and providing an output and it's not paying back the original creators of that yeah which is gonna prove a larger problem than we think in the next year because no one would want to go into these fields if they are not protected by regulations i mean that that is an interesting you know conundrum i guess is like we always see that you know your legal regulation always comes after the

[00:53:58] fact particularly when it comes to technology because it's always a bit challenging like say to kind of predict where things are going to go i mean i will not dwell on that too often but i mean i guess in that in that kind of ghibli example you know i guess if i cast my mind back before chat gpt came along you know we've always had people who would copy other people's art or copy people's styles and things like that and that has been protected in various ways because we've got ip protections and other things they do sit around that equally you know you know a ghibli original

[00:54:26] will sell for way more than a copy would would ever do and and and so the key thing is being able to identify is that an original or is that a copy and i hopefully at the moment at least with the whole chat gpt thing it's fairly obvious where it is being generated by chat gpt um i guess the conversation here then moves us on a little bit to sort of like thinking about ethics and and you know um you're both icaw members as well so that i know they're really hot on on on ethics and the impact on the

[00:54:53] profession but but can you um you sort of talk around you know what are the challenges particularly for accountants bookkeepers maybe who want to make more use of chat gpt or something similar what what implications do we need to consider as a consequence well the regulation is the most important thing for humanity at this moment regulation of ai unfortunately we are moving too slow and we're not moving all together i think that's a larger issue than we're moving slow because we all

[00:55:22] have to agree between us on how to regulate the ai i mean if the european union is the first one that applied the regulation they firstly establish ai act but the european union is the only one that applied the regulation until now and it will have no impact if the rest of the world decides that i'm not gonna apply regulation because i'm gonna be left behind and uh i won't have innovation within my country to develop ai because regulation would take me back

[00:55:52] so it does nothing by just one uh institution applying regulation we all have to come together and agree on something and i'm sure you understand how hard that is it never happened in the past of history unfortunately and we would need to definitely find a way because it's ai is still an infant and it's now the time to bring it up with the right morals with the right ethics

[00:56:20] and we're not doing that we're leaving it to learn out of public data which may be wrong may be biased and maybe it will be brought up to become unethical and harmless in a way possibly in the future and we have seen that haven't we we've seen some models having to be pulled and obviously chat gpt have just made an update or we were all back an update because the model was too sycophantic and

[00:56:45] kind of you know gave you praise around you know potentially making your unethical decisions in some instances um but but whose responsibility is that because i mean i guess you know i i can understand the kind of the the moral debate between um regulation versus innovation i don't i don't think actually there's ever been a time where regulation has you inhibited innovation because i think innovators always innovate irrespective of the frameworks that are working but who's going to

[00:57:15] be who's going to be responsible for feeding information and saying this is reliable and this isn't reliable i mean we saw that we saw this challenge a few years ago when um i think was it google were labeling news sites as being reliable or not and there was a huge amount of pushback on that to the point where it got it got completely quashed i mean are we going to have to establish some new societal norms to to allow this to happen um again no answer for that but we can brainstorm if

[00:57:42] you want i mean it's it's all new it's all new to me it's all new to you it's all new to humanity in general we don't have the answer of who should be responsible and it's one of the questions about when they when organ when countries trying to set up their own regulation one of the questions is whose responsibility who should be accountable for the outcome of these models i mean if i try to trick

[00:58:05] chat gpt to give me a something unethical as an answer and i've managed to do it is that my responsibility is that the model's responsibility and the programmers behind that designed it and allowed it to do that is that the data's responsibility where did we find the data so it's a very hard question to answer and i do not have i'm sorry about that i do not have an answer uh but i do agree with the question and i believe it's very important that we answer that question

[00:58:34] who should be responsible for the good outcome because we do want a good outcome and and not a non-harmful outcome for all llms yeah i i always think i mean i think this um um you know this llm debate in particular or having recently really um really highlights the kind of tech fallacy that we see in in society in general and and you can reflect this onto something else which is like the self-driving cars you know thing at the moment where um i don't know if you've heard this monica but

[00:59:02] basically you people expect self-driving cars to never have accidents because it's technology so it should be perfect um and then and then obviously they do have accidents occasionally and more often than not the accidents that they've had have been caused by the human being in the vehicle not paying attention or it's not been trained on the right information or whatever and it's always a human failing not a technology failing that that's that's arisen more often than not um and and yet like i say we expect it to be perfect and we're still having the same debate now it's like who's

[00:59:31] responsible for the accident you know in terms of who just who's who's insuring this thing who is is it the manufacturer that should be responsible is it the person the vehicle that's responsible and and we still don't have answers answers there do we um moving moving on then oh sorry if i may add about that no we don't we don't what uh what the ai act did the european regulation on ai they uh

[00:59:55] they asked for high risk models ai models that fall under the category of high risk to have a human oversight so so a human uh a person to be to be able to validate that the outcome of the ai the decision is validated by a human so it's adding in a way responsibility to the back to the human yeah which i guess kind of makes sense right you know in in the grand scheme of things i think particularly in a

[01:00:23] professional services world you know if i think about um you the things that we're responsible for when we provide our professional services you know we're providing an opinion or we're signing an audit report or signing an account report or something else you know it always falls back to the um the individual to apply their own judgment but i guess moving on from that and and you know conscious of time as well is that sam you we talked about capability of these models and we talked about

[01:00:50] skills and things like that and you know one of the one of the things i read fairly recently sort of said that the the um you know the task capability of of large language models is in doubling in capability every seven months um and you talked about agentic ai which is really like the big hot topic that we're seeing at the moment but you know what what is the conceivable risk or what is the real you know the real issue for you know not just professionals but other people in business you know around around

[01:01:16] skills you know i read uh linkedin provided an article which touched on the fact that they expect 70 percent of skills currently in use by by people working in an office environment or professional environment to be gone in the next five years as a consequence of ai so that's some big impacts that we've got to consider right yes and uh it is predicted i mean jobs have been changing

[01:01:40] throughout time so jobs that we had like 100 years ago do not exist now i mean a horse changing thing something whatever does not exist now not my job was not did not exist like 10 years ago as well so jobs have been reinvented throughout time so that's not an issue because we are able to reinvent ourselves anyway what it is predicted however is the speed that is going to happen now we are expected

[01:02:11] to reinvent ourselves to change positions five times during our lifetime because of the rate of change that things are going to be we're going to be seeing in the next years so the the skill that we need to build is to constantly learn and not be found my job now when i sit back everything's perfect because it's things are going to be moving whether you like it or not so the most important skill is to be adaptable

[01:02:39] to be able to constantly learn and adapt to new situations and not not be afraid of uh of change oh that's amazing i mean i love that i mean i i love the fact that you know we're talking about you um you know when you made your introduction you talked about the fact that you're curious and you're interested in in the opportunities and the things that fascinated you and you were trying to

[01:03:02] find find somewhere that kind of you um allowed you to express yourself and and and you express that curiosity and interest and now of course we're talking about being adaptable and being able to change and you know all of these key skills you know i i mean i wonder if are these are these key skills or these traits you know are these things that um you know you can learn or do you just have them you know naturally you can definitely develop i mean some of our inherent i mean i was fighting

[01:03:28] with my mom when i was a child because i was taking up a hobby and then after a month i was like now i'm bored now of playing the guitar i want to learn something else so she was like no you need to stick to something you need to learn this thing and become good at it but i was always trying to learn something new so i believe i was born with this curiosity and these fast changing things which in another decade might not benefit me so much but now it much i match with the time so that's good for me

[01:03:58] but you can definitely you can definitely develop it you uh by i mean there's lots of things you can do we now have more information than humans have ever information i mean if if i remember the largest i mean do you know the library of alexandria yeah it was the yeah it was the largest library it was the largest source of information that people had at the time now this this thing that we're holding our

[01:04:24] phone holds more information than that library uh had that than people had access to information so we do have the information we i mean you can read you can watch youtube videos you can watch series you can see podcasts you have the information in all formats and whatever suits you you should be

[01:04:46] uh constantly learning and constantly following what's going on not specifically about technology on how ai is moving but specifically for your industry in the industry that you're interested in let's say i'm an artist then let's say i'm a painter then you constantly need me to learn what's the updates what's going on with painting well how should i reinvent myself should i learn about what nfts are

[01:05:11] should i find out how i'm going to have be sustainable in the future so whatever industry you're doing ai is going to impact that industry so you need to be in front of the game and not following behind so that's yeah that's really important amazing well thank you monica for your time um i mean i think probably talk about this for like three or four hours if not longer um and maybe we'll just get

[01:05:37] you back on another point and we'll talk a bit more about this or or when we get to the doomsday scenario where no no accountants are left and we can reflect on them on who's left standing and what roles are still working for us but um thank you very much for your time uh it's always great to speak to you monica and no doubt i'll catch you on the next uh icw meeting but um um beforehand if anyone wants to get hold of you how what's easiest way to contact you if they want to learn more or maybe get a bit more involved in what you're doing out in cyprus yeah you you can follow me on linkedin

[01:06:08] um which i usually reply fast so and we can take it from there if you're interested to talk more about talk to me about ai and talk to me about how we should apply ethics on ai amazing amazing okay thank you very much monica that was awesome thank you john and that brings us to another end of our podcast so thank you for listening if

[01:06:32] you didn't manage to uh to catch us at uh account text then look be aware we're going to be to bring other events to your ears and we'll make some notifications about that releases some announcements about that in the coming weeks and months um but we do always like to hear from you please do reach out to us if you've got any comments queries or just to make us aware of your product and what it's been doing we like to keep our ears to the ground we like to make sure that we are at the forefront of all technology so that we can inform you and make sure that you know what's

[01:07:02] going on out there please do like follow and also leave a review and we like to hear your thoughts and feelings and any improvements we can make to what we're doing um and yeah we'll catch you on the next episode thanks for listening