No Accounting for Taste ep215: AI reality check for tax tech

Speaker A: Shall we begin? Let's begin now. Hello and welcome to the latest episode of no Accounting for Taste, with me, Matthew Ord, and the familiar voice of Accounting Web's technology editor, Tom Herbert.

Speaker B: Hi, Matt. Hi, everyone.

Speaker A: It's just the two of us again today with Richard out and about in London. And while we would typically mourn his not being here, who will instead use it as a flawless segue as Tom, you were there yourself recently.

Speaker B: Uh, yeah, literally yesterday, as we record. So, uh, yeah, absolutely remarkable scenes. Um, on the train back, I was waiting in Paddington Platform two, if you're interested. Uh, platform one's the one with Paddington, the statue on. Anyway, I got a window seat, I look down and on the track is a very well fed fox with a rat in its mouth. You know, just trotting along. I sort of scrambled for my phone to try and get a picture before I could, uh, you know, because no one's going to ever believe me that this thing's trotting along. But unfortunately he ran off to his foxy hidey hole before I could snap, uh, a snap, an image of him. But, uh, yeah, quite the, uh, quite the picture.

Speaker A: You'd typically be disappointed if you weren't on the actual Paddington platform, but it sounds like you've looked out in a. Well, looked out. Maybe the wrong turn of phrase.

Speaker B: Yeah, it was pretty gory. I was discussing this with child too. You know, should we be happy for the fox that it's got its meal or less happy for the rat, maybe. But, um, there you go. That's life.

Speaker A: It is, it is. I think if David Attenborough ever runs out of any kind of funding, uh, this is where he'll find himself. Just sat on trains looking at wildlife in the London area. Hopefully not the highlight of your London trip though, tomorrow.

Speaker B: No, absolutely not. No. This was. Yeah, it's been a busy event season. I've been to lots, um, up and down the country and abroad. But, uh, yeah, this was Tax Tech Live at the London Science Museum. So, yeah, lots, lots going on from there. I guess the biggest news was that it's, I mean, it's an event from Tax Systems, but Tax Systems is no longer Tax Systems as they've rebranded to Alpha Tax. Obviously a big beast of UK corporate tax, uh, FTSE, 100 companies, big firms, etc. They've been around for 35 years, but I think Alpha Tax was always the product, whereas Tax Systems was the company name. I guess, similar to us. I mean, uh, our, uh, parent company's called Sift, but if I wander around the show floor saying, yeah, I work for Sift. Then you know, everyone would sort of stare at you blankly. Whereas accounting, uh, web is uh, known and loved, you know, I think it's a decision that makes sense. Chatting to Russell Gammon from Alpha Tax, as I should say, he said, you know, there's still a lot of, still a lot of people. Even though Tax Systems is the name, a lot of people will just say Alpha Tax because that's what they log into in the mornings to use. So yeah, all seem very sensible.

Speaker A: Yeah, brilliant. And what were people saying at the conference then about how are they using AI, uh, Tom? Cause I imagine that was on everyone's agenda.

Speaker B: Yes, yeah, I mean they were quite sensible in the way that they stacked the conference. You know, there was a bit about their cloud migration. Obviously it's a big desktop beast back in the day, new product and integrations. But you know, they left the bit about artificial intelligence till a bit later on. It was actually quite heartening to be in a room with sort of senior experienced tax tech professionals. Um, and for us all to be sort of looking at each other saying, so yeah, what are you doing with AI? Just a huge amount of noise. I get it. Like no one likes to not conform. You know, everyone wants to say, oh, I'm using AI to revolutionize my workflow and I've got 120 hours a week back, you know, and all the rest of it. But I think once someone breaks the ice and says, well I'm using it for this little bit here or I'm testing this bit there, everyone's just sort of saying oh, phew, okay, great, you know, I'm not wildly behind, I'm not some kind of Luddite who uh, is just denying it for the sake of it. So yeah, in terms of how they're using it to answer um, your actual question, there are gains, sure. They're kind of marginal though, you know, I think Tax systems, Alpha Tax, sorry, slip one there, I guess things like trial balance mapping, answering questions, digging out sources. They acquired this Belgian tax automation assistant called Lock Tax back in October last year and that's quite handy for answering some of those cross border tax questions. Having all your information in one place, coming up with like I said, sort of sources to questions you've asked it as it were. So yeah, but as I say, I think what's been interesting, uh, not only about the Alpha Tax event, but also just, just event seasons in general, the fog is slightly lifting. I mean a couple of years ago it felt like this new generation of AI was a solution looking for a problem in accountancy. I mean, sure, it could do some neat things, but how much could it really help? What specific areas? I think it appears to be now slowly coalescing around information, whether that's getting documents from clients, flagging missing information, sorting it, tagging it, reconciling it against bank statements. So a lot of that sort of messy work, trying to automate some of that. One of the more interesting solutions that they had an integration section where, I mean, they talked about their integration with caseware. So two big beasts of the compliance world there. But, uh, right at the beginning of their journey is something called Platform 3, which essentially automatically applies IXBRL tags to updated financial documents. That's using machine learning. It flags where complex judgments are needed by the user. Ultimately it produces like a compliant ixbrl output file ready for submission with a full audit trail. So, I mean, given what a pain ixbrl tagging can be, it feels genuinely useful. And there are a lot of the sort of smaller, uh, examples like this where uh, it's sort of chipping away without being, you know, completely revolutionary.

Speaker A: So where do you think this leaves us in terms of the shape, the profession and the training that goes with

Speaker B: it still feels very nascent. Um, there's a. Given the noise, given the column inches devoted to it, it feels wildly out of proportion, to be perfectly honest with you. But as I say, it feels like vendors are so desperate to talk about it because they don't want to be left behind and um, equally just dragging professionals along in their wake. It feels like there are a couple of directions. I mentioned Alpha Tax's Russell Gammon there. He used the example of radiologists. So 10 years ago, AI, uh, I guess probably sort of machine learning scientists back then were talking about how AI was going to, just going to completely obliterate the job of radiologists. So taking X rays or scans or whatever, and just because of the vast amount of data that it could train from just being able to quickly pick out what was wrong with someone and flag it. But actually compared to 10 years ago, now they're actually more radiologists. And actually a lot of their work has been streamlined, they are now very much viewed as sort of being able to use their expert judgment on things. And I think that the example being given there of the sort of tax professional being able to apply judgment based on context, I think is fairly relevant. A couple of people I chatted to were sort of perhaps saying maybe we will need fewer trainees you know, the classic pyramid structure at uh, those larger firms where you know, the small number of leaders at the top, but they need a huge volume of new entrants or trainees or sort of lower grade members of staff to sort of churn out the work. Maybe that pyramid might sort of start to uh, reduce or become a diamond or what have you, you know, and it's been interesting just in my position from a technology standpoint, the number of new tech vendors started up by people leaving the big four. It's completely unscientific, but it appears to me you get to a certain level and you perhaps get your head turned by the world of tech, think that you find a problem that needs solving, you hook up with a developer or an AI scientist and build a product and email uh, me about it. Yeah, it's interesting. I was just reading a story this morning about training. I think perhaps one of the concerns that people have is around the fact that with a lot of this lower level work being automated, how were trainees going to build up their experience of these real world scenarios, uh, in order to become experts and apply their judgments. And uh, I was just reading a Business Insider that KPMG's testing a uh, sort of simulation tool designed to help employees build those skills. Um, you know, that you would have gained previously through years of pencil pushing on, uh, a bookkeeping standpoint. You've got products like Heymonica that do that as well. So yeah, I guess it will be different. Will that be a substitute for real world experience? It's impossible to say at the moment. But yeah, I think we will come to a crunch point, I think relatively soon. Comparing it to the dot com boom, you know, the sort of first wave of the Internet I think is incorrect in a lot of ways because the unit economics of the Internet of E commerce were always excellent. You know, it's just that you sell and you keep selling and the more you sell the more profit you get. Whereas actually, I mean particularly at the moment, the more people use AI, the more it's costing these people because it's being so subsidized. We are going to come to a crunch. We are going to come to a point where they have to raise the prices sort of exponentially. You know, we're talking double or, or triple. A lot of these companies using it, they aren't charities, they're going to be using this AI because it's more efficient, because it's cheaper than using workers. But once the price levels sort of stabilize as it were, is that still going to be the case. So yeah, I think we will come to a crunch point, but how soon that's going to be is impossible to say.

Speaker A: I want to just return to an earlier point you made, Tom, where you said it was heartening to be in a room where these kind of senior and experienced tax tech professionals were all not quite sure what they were going to do with AI and that that was a bit of a phenomenon. Do you think that might not become the norm by any means, but do you think people might start to be a bit more open about the fact that they're just not quite sure where to take it?

Speaker B: I hope so. I hope there is. I mean one of the great things about the profession is how collegiate it is is how people wandering around and just sharing examples of best practice or what didn't work for them. And um, for whatever reason over the last couple of years that hasn't quite been the case with AI. Whereas now I think people are reaching a sort of point of understanding the limitations of the technology or some of the applications. And so I think that we will, yeah, hopefully be able to sort of get to a point where people are having open and honest discussions about it all.

Speaker A: Here's hoping, like we said. Yeah, he's hoping that it needs a little, a little bit more of that, doesn't it? No one pretending that they're kind of flying and are a full blown AI native when really there's all of us just still trying to figure it out.

Speaker B: I think mandating it isn't the answer. You know, I mentioned KPMG earlier. I think over in the US there was a report that essentially KPMG is forcing its staff to use AI and they've got to log their AI, uh, uses. You know, you've got to use it at least 75% of your working days. So if you're on a four day week, you've got to have used it for three days. But essentially what they're doing, from the sounds of it, is a giant spreadsheet and people are just generating fluff, generating just prompts just to say they've used it and then moving on. You know, that's, that's not the answer. No, you know, education, education is the answer, uh, testing. And I think that's why in a lot of ways, I think I posited on this before that actually smaller firms, in this initial wave, anyway, smaller firms have a slightly better uh, footing to sort of trial these things. They can be a bit more flexible, sort of, you know, take an afternoon to see if something works or not without fear of reprisals from, uh, IT managers.

Speaker A: Yes, agreed, agreed. It seems like you're right, an interesting tactic to try and get people to use it, but have this kind of lingering presence of, well, if you don't, who knows? Just seems a bit odd. No one that's going to be encouraged to do it that way, surely.

Speaker B: Yeah, tech is, you know, uh, generally adopted from the bottom and sort of rises up forcing people to use this stuff. Just, it just seems mad. Like, you know, the idea people, I guess in the early days of the Internet, you know, workers would literally bring in stick drives or sort of find browser tools to sort of help them. It was the sort of wile e coyotes on the little train tracks of just IT managers sort of frantically laying down, you know, what people could do, um, in front of them, as it were. So yeah, it's, it's kind of fascinating, but I guess when the, the likes of your, your Lotus Notes and um, your Microsoft, uh, Office, uh, all came in there, there was a sort of uneasy truth truce, should I say, around that. But yeah, it's a fascinating world at the moment just, just seeing how it's all going to play out.

Speaker A: It is indeed. It is indeed. I'm sure we'll pick this up next time around as well. Tom, thank you as ever for the insight. I want to talk about the sentiment of finance leaders with EY having released their global DNA of the CFO survey this morning at the time of recording covers a little bit of everything. But the key take for me is that CFO ambition is outpacing execution. Now we've touched on that perception performance gap a couple of episodes ago and it's come to light again here, which kind of proves that it is this real genuine problem with 60% of CFOs saying they want to lead on value creation, but only 26% actually leading value creation discussions and 25% leading vital investment decisions, which isn't very much at all really. There remains a hurdle for CFOs in translating those aspirations into influence, with only 27% saying that their organization see the finance function as a key partner when it comes to value creation. There's a line in the survey that says without stronger ownership, CFOs risk remaining reactive rather than driving enterprise value. And therein lies the big concern. There's a lot less value in being reactive. In a lot of ways it's almost the case that being reactive is actually to be too late. I'm sure you'll be Able to back me up on this one, Tom. Remember that famous saying from Paolo Maldini? He says, when I've got to make a tackle, I've already made a mistake. Feels a little bit like that.

Speaker B: Oh, uh, friend of the show, Paolo. Blimey. I wasn't expecting that kidness. Yeah.

Speaker A: Uh, but it's a little bit like that. You can't predict everything, of course, but the draw of the modern CFO who's making a difference is their ability to get and stay ahead of these things, be it through forecasting or knowing the market really well. Remaining reactive just loses that extra something that I think the modern CFO would love to have. A lot of them feel like they should have, and rightly so, but it's, there's, there's just that, that missing ingredient there. Modern. Two thirds surveyed also spoke of new skills and leadership styles being required. But people in culture leadership ranked only sixth when it came to development priorities. So again, there's that big thing holding them back to being disconnect. It's just the expectations. What they want, what they need just isn't quite matching up. Looking at what comes next, the study notes that the future of finance will not be determined by talent or technology alone, but by how effectively CFOs bring them together to create that value. So what we've got, Tom, is a finance leader sitting with a lot of the right tools around them. But be it through businesses not seeing them as a key partner or teams unable to adapt to new technologies as fast as they'd like, some of the struggles revolve around just being able to bring it all together. They've got the pieces for the puzzle, but connecting them is just proving to be a bit of a bind. It's a, it's a funny one really.

Speaker B: Old habits die hard. There is that sort of, uh, perception from, uh, the very top level of seeing finance as a cost center rather than actually a value creator. You know, it still is, you know, Nigel in the corner office with his spreadsheets rather than someone. The accounting system literally at the heart of the business with all the data. And I guess we talk about proactive and reactive. I guess it's up to the finance leader to use these new systems, you know, to use any tools at their disposal in order to demonstrate that value, pulling things together and, uh, sort of bringing them to board meetings or management packs or whatever else they're using it to produce and just going, you know, above and beyond and just saying, well, look, I've noticed X or have you seen how Y is tracking or something like that. It's actually because people are busy. They're going to be looking out for it themselves or they won't be able to look out for things like that themselves. So yeah, I think it's a little bit chicken and egg. Who's going to drive this, uh, change here? But um, I think there's plenty of positives to be gained from a lot of these things.

Speaker A: Agreed. And I think it's that age old thing as well of finance folk aren't very good at shouting about what they're good at. Not very good at sitting up and saying, hey, we can do this, we can help, we can make a change. They feel like they can and they absolutely can. But I think that needs to be a bit of a mindset change, a bit of a behavior change from them to say, put this data in front of people, go along to those board meetings and be able to kind of verbally and noticeably make uh, a, make a difference. It's a, it's a bit of a hurdle, isn't it?

Speaker B: It doesn't need to be transformative. You know, even one stat or something you've noticed from a particular department can be anything but, but, but just putting those building blocks in place.

Speaker A: Yeah. A bit at a time. A bit at a time. You're very right. Well, let's see ourselves out with a return to the audit beat. It feels like it was front and center at the start of the year when there was some optimism from me and me alone, it seems with my garbage predictions at the end of last year, it seems things were possibly moving in the right direction before plans for audit reform were, um, unceremoniously pulled. Since then it's all gone a little bit quiet. But a story landed this week written by producer Molly, which has reshone a light on, um, why trust in audit is worryingly, but understandably quite low. BDO was slapped with a 2 million pound financial penalty, reduced by 5% for exceptional cooperation and then discounted by a further 30% for admissions and early settlement to make it 1.33 million. After the FRC found numerous and pervasive breaches in its audit of construction and consultancy group NMCN M PLC. Now NMCN was incorporated as, uh, North Midland Construction Co. Ltd. In 1946 and was first listed on the London stock exchange in 1982. So there's a storied history there. We love a bit of operated as a construction and consultancy services group delivering major building and national infrastructure projects across the UK. However, it entered administration in October 2021 with the FRC notice concerning both BDO and Audit Engagement partner Geraint Jones over the statutory audit of MNCN's financial statements for the year ended 31 December 2019. The FRC said the breaches were significantly serious, with the failings mainly related to BDO's audit work of NMCN's long term contracts. The audit team had already identified significant risks around revenue and profit recognition on those contracts, as well as the recoverability of contract assets, trade receivables and retentions. But the FRC found that the audit work carried out in response to those risks was not good enough. As a result, BDO and Jones failed to obtain reasonable assurance over whether the financial statements as a whole were free from material misstatement. The FRC also found failings in the audit of Going Concern, including a failure to plan and perform the audit with professional skepticism. Now, uh, while these breaches were serious, we should say that FRC didn't suggest that they were intentional, dishonest, deliberate or reckless. And it also didn't claim that NMCN's financial statements were misstated. Having said that, it brings us back to the long term problem of audit work just not being up to standard. Far from it. In fact, it does nothing to restore trust. Tom. And with audit reform Bill in the bin, the question of what happens next is just kind of hanging there a little bit, isn't it?

Speaker B: Indeed, yeah. Uh, it's an old favourite. You know, the cycle of life continues. You know, we have a few corporate collapses that mysteriously affects the taxpayer, government promises to do something about it, produces a fairly, uh, strict set of proposals. It, uh, ultimately gets watered down or in this case, you know, governments change and uh, it's no longer a priority in a financial crisis. And then uh, something might change. In this case, nothing's changed and so, you know, we're kind of just waiting for the next round of collapses really, I guess, until um, something's done about it. Did you say Geraint Jones? That's the name there. I'm really hoping it's not former England wicketkeeper Geraint Jones. It would ruin my uh, ruined my image of him taking that, that flying catch on the leg side of uh, Steve Harmerson in the 2005 Ashes. But uh, I'm hoping it's not him.

Speaker A: Yeah, Let me confidently say for the purposes of the show, it is absolutely not that Geraint Jones for sure.

Speaker B: Thank goodness.

Speaker A: But yeah, you're right, we are just kind of sat here waiting for what comes next. And I remember at the time that the audit bill got scrapped, there was this resilience from the profession who kind of said, well, that's whatever it's happens. We kind of knew it was going to come. We'll take this on and we'll sort it out. But to a point, there's only so much they can do. And we haven't seen any progress at all, really, in the, in the five, six months since then. So you're right. We just, we just sat here waiting on, on what happens next.

Speaker B: Yeah, we did a really interesting, really interesting webinar, uh, on, um, tech and the future of audit with our friends at Case, where fairly recently, well worth checking it out. I think it's in like the resources section on the site. If you go to accountingweb.co.uk, it's, it's, you'll find it. But, uh, you know, I think there's, there's so many new tools coming into audit. Rather than sampling, you know, some of these AI tools can look at entire data sets and flag things that needed to be flagged, as it were, and cut out a, uh, lot of the grind associated with audits. So, yeah, I think a, uh, silver lining of all this might be that as and when audit firms, um, start to adopt these new tools, perhaps we can take a fresh look at, um, the audit profession and look at how, uh, best to regulate it in a new era of technology.

Speaker A: Agreed, agreed. I think, yeah, a clean slate is probably as good as any at this point because if anyone's able to keep track of where we even are, it would be impressive. So, yeah, I think it sounds like a smart plan, Tom. Sounds like a smart plan because it's one of those, if you just pick it up now, you've got everything that's come before it and, oh, it's just

Speaker B: such a mess, you end up with MCD essentially, you know, a tax digitization program for the 2010s.

Speaker A: Yes, agreed. And a finer note as any to, uh, end the episode on. Thank you, Tom. Thank you very much indeed. That's all we've got time for this week. Join us next time for a look at all the big goings on. But until then, take care.