How Taxd Is Building AI Tax Automation With Humans in the Loop

[00:00:00] Do you need AI agents that you can trust? Well, with an AI data layer providing real-time connection within your data platforms, you can trust your agents to provide accurate solutions. So, scale your business by trusting your agentic AI accurately getting the work done for you. Trust its capabilities with Denodo. And you can do that by simply visiting denodo.com to learn more.

[00:00:30] Would you trust AI to prepare your taxes? Now, before your answer, remember that one confident mix-up between HMRC or the IRS could turn a clever shortcut into a very long, stressful afternoon. So, today on AI at Work, I'm joined by one of the co-founders of a company called Taxd. And they are a digital filing platform for individuals and small businesses.

[00:00:58] And my guest will explain today why routine tax compliance could become autonomous, but advice, context and human judgment will continue to matter and arguably never be more important. And we'll also talk about how taxed is using AI to review large volumes of transactions. And why a software-only approach will inevitably struggle in a business built on personal data and trust.

[00:01:27] And how real-time reporting is already changing what taxpayers expect from accountants and technology. So, if you are introducing AI into a regulated industry, building a startup, or simply wondering whether a chatbot could be trusted with your tax return, whether AI can be trusted with your tax return, or interested in how AI is transforming this particular area of the business,

[00:01:55] I'm hoping that today's conversation today will give you a useful line between smart automation and expensive overconfidence. But let's get started. Let me introduce you to my guest right now. So, thank you for joining me on the podcast today. Can you tell everyone listening a little about who you are and what you do? Thanks, Neil. It's great to be on the podcast. My name's Arj. I'm one of the founders of Taxd. We are a digital tax filing service platform.

[00:02:25] We help individuals and small businesses to file their taxes online as efficiently as possible. When I describe that to some people, they think we're the tax man. We're definitely not the tax man. We're trying to help people stay efficient and compliant at the same time. And my background is from the PwC. So, big for large accounting, large tax world. I did my tax exams there and met my co-founder as well. And then we started Taxd over COVID.

[00:02:52] We're one of those COVID-born businesses through the lockdowns. And we started primarily because we saw how old school things were being done with offshoring to India and that side of things. And that's why we wanted to start Taxd was to create something a little bit more efficient, something that uses technology and not just moving things overseas

[00:03:18] and actually creating automation that helps people be more efficient and compliant. I love that. And to dig a little bit deeper on that, I'm a sucker for a good origin story. So, what was it that first attracted you to the world of tax? And what problem in the tax ecosystem just made you stop and think, this has got to change? And where do you think maybe traditional tax services are still failing modern businesses today, especially now AI is in the workplace? Interesting.

[00:03:47] Just on that first point, what attracted me to the world of tax was honestly, I did a school leaver scheme. So, I didn't go to university. So, I was just seeing what apprenticeships were out there. And PwC was there. So, I was like, oh, that seems like a good option. So, it kind of fell into the world of tax. There wasn't really like, I didn't have a revolution to try and change the kind of tax world when I was younger.

[00:04:16] It was just falling into it. But then I think with regards to seeing how things were working at these big firms, there is that niche specialty and expertise. And that's what we found, myself and my co-founder, Eamon, we both worked in expat tax in particular. So, we used to help individuals that are coming in and out of the UK. And it's a very niche part of tax. Your high street accountant wouldn't really have that kind of knowledge either.

[00:04:43] It's looking at things like what happens if you move to Spain? You might pay tax in two places depending on your income, that sort of thing. Now, what we then found was, and I don't think this problem is stuck with tax. This is across the board with the partnership model. So, same in law firms, in general, professional services.

[00:05:08] Where there's a partnership model, the model often incentivizes to maximize profits on a year-on-year basis and not so much to invest for the future. And I think that's what we saw where investment in tech wasn't really a priority, but reducing costs on the bottom line was. So, that's where kind of offshoring became such a big thing, where it was easy to actually create more margin from clients.

[00:05:37] But then also, yeah, the tech investment really wasn't there. I think that's some of the key things that we saw in the tax world. But I see this across professional services, really, where that upfront investment isn't really there. So, I think that's what we wanted to really tackle was to say, hey, actually, we can build it. And it's kind of funny, when we started off as part of the origin story, we first actually went to some of the partners and we said,

[00:06:07] hey, we can build this in-house. And they said to us, no, we have a transformation team that's taking care of it. So, don't worry, essentially. And then we were like, all right, well, we can just do it ourselves. So, it was a bit of a push because I think it was a push to take that risk and to set things off on our own first. What a great story there. And I think we forget just how far we've come. I mean, you mentioned the tech investment just wasn't there.

[00:06:36] And if we go back, what, six years ago, even just working from home was just for the privileged few. And now, of course, everyone was working from home overnight. Now, it's gone into hybrid working. And in the last three years, we've seen AI enter the mainstream and now dominating the workplace. And I think every person listening to this podcast is trying to work out what could AI improve inside their workplace. So, from your work point of view here, how is AI transforming the tax workflow with what you're doing?

[00:07:06] And where does human judgment still remain essential? Because I think that's something that everyone's grappling with right now as well. Yeah, 100%. And I think that's, we're going to get to a point, in my opinion, where things like compliance will be autonomous. So, tax compliance, where all the data is already out there. There's not much AI that really needs to make a judgment call. It's pretty much taking data from different sources, putting it onto a tax return.

[00:07:34] That compliance element is what we are seeing becoming more autonomous. But the advisory side of things, I think that's where AI will be powering that element of it. That's where there will be more of a judgment call and humans in the loop, the relevant tax advisors supporting that process. So, from a tax perspective in particular, where we're seeing it being deployed is AI supporting with things like reading through large amount of transactions.

[00:08:02] So, when you think of a sole trader, a landlord, a business owner, often there's hundreds of transactions to support, produce that tax return. And that's the sort of thing that AI can help with. But for a simple tax return, it's just going to be, it's not really AI, it's just the rules in terms of trying to take data from different sources. And for most people nowadays, all that data lives online. I know it's either in places like HMRC's own portal, it could be in trading accounts,

[00:08:31] and kind of brokerages, crypto, etc. Just about connecting those dots. And where people might have lots of transactions, things to analyze, that's where AI can help. And then the human in the loop might still be reserved for certain areas, which is still super important, where that extra bit of judgment is required. I think that part of it will still need to be kept as a human,

[00:08:57] where AI may not be able to do 100% of things. Yeah, I completely agree. And there's so much excitement around AI, in particular around everything professional services. And of course, there's also plenty of misunderstanding, noise, and even hype. So I've got to ask you, in a world where everyone sees AI as the hammer to every single nail out there. What is one misconception about AI and tax technology that you would actually challenge?

[00:09:27] Maybe we could lay it to rest today. Yeah. So I think quite an interesting one that we get from clients is clients who have used AI as their tax advisor. And then they come to us and then are questioning quite simple things where HMRC has said it should be done this way. But the AI might have taken some information from the IRS or tax authorities around the world, right? And then it's telling an individual to do things a different way.

[00:09:56] And I think that is essential where you are using AI to almost have that human to check it. And I think that's why you have experts, advisors, et cetera, who may have done this thing for years and have the experience in that way. I think that's where there's almost that misconception that AI can complete these tasks because it's logic-based. And actually, when it comes to tax,

[00:10:26] although a lot of it is logic, there is still an element of, one, there is a judgment piece. And two, there are times where the logic overlaps. So I think that's where a human still needs to come together and actually review the situation in form. Yeah, I'm glad you mentioned that because I think tax is such a highly regulated industry and rules do vary significantly between every jurisdiction and every continent of there.

[00:10:54] We've got one AI provider that's just hoovering it all up and giving you generic answers. So yeah, the importance of humans is so important there. But what has surprised you most about building and scaling technology in such a complex regulatory environment that varies from region to region? Yeah, I think one of the key things that we found in regulated and quite a personal industry as well, tax is super personal as it involves a lot around

[00:11:24] people's earnings, right? How much people earn, how much people spend on things, their investments. So we found just a pure tech approach doesn't work for everyone. It might work for some early adopters who are more comfortable with, hey, like tech have my data. Whereas I think for us, what we've really tried to do is be a bit of a hybrid. So we've got software, we've got AI, but you've also got humans behind the scenes.

[00:11:53] So if you want to speak to someone, you can. And I think that's how we are getting around and supporting people in a bit more of a tricky industry, I'd call it, where that trust element is so key. I think that really does involve trying to do things that don't scale as well. Like we had lots of calls with clients at first, really trying to build that trust and seeing how do we get to build the trust within our own user experience from a tech perspective,

[00:12:21] but then also just enabling individuals that if they do need to speak to someone, always providing that option there. And as businesses inevitably move towards real-time finance and faster decision-making, where does your work here, how does that fit into the broader financial ecosystem alongside banks, accounting platforms, payroll providers, and so many other business tools, where do you guys fit here? It's a good question in terms of like the real-time tax and finance. HMRC are definitely moving towards this.

[00:12:51] You've probably heard of MTD. If you're a listener's perspective, that's a new rule that came out just the last couple of months, which where individuals, sole traders, landlords have to file now five times a year. So four quarterly updates throughout the year. So much more real-time reporting than the traditional wait till year end and then file. I think that's what we're seeing the tax authorities move more towards. It's already the case in places like Spain as well, but I think we'll see it across the board.

[00:13:20] And in terms of where we sit, third-party software, I think is going to become key to bridge the gap between that tax authority wanting real-time reporting and then actually getting the individuals to be able to report in real-time. There needs to be that third-party software to facilitate that kind of bridge between the two. And where we sit in amongst the banks, payroll, et cetera, et cetera,

[00:13:49] we taxed position ourselves as a hybrid between that accountant and software. So if you just want software to run payroll or you just want software to file your taxes or do bookkeeping, we've got that. But also if you do need that accountant for some advice, or as you're using the software, you know that you want to pick up the phone, speak to a real person, you can do that as well. So I think that's where we kind of position ourselves quite nicely as being able to cover

[00:14:18] almost the same services they traditional accountant would, but doing it in a scalable and for the individual a more affordable way too. And I think one of the biggest challenges for any new solution is always going to be not the technology very often, but adoption, company culture. So a question I'd love to ask you here, what has been harder, building the technology itself for convincing people, organisations and teams to change the way

[00:14:47] that they've traditionally approached tax? And hey, this is the way we've always done it. And from that exercise, what have you learned about maybe trying to drive behavioural change? Yeah, I think for us, the behavioural change that we've really seen is there's a stark difference, I think, between someone filing their personal taxes and then for their companies. Yeah. And that's what we, when we first launched, we were only in the personal tax space. I think people were already used

[00:15:16] to filing their taxes themselves. It was something that people were more confident in doing. When we shifted over to supporting the company side of it, getting someone to trust us with their company accounts, which is also information that is shown to the public on Companies House, was, I think, a different kettle of fish completely. And people, I think, kind of business owners have been told that they need an accountant for that kind of thing.

[00:15:46] They need someone else to handle it. So I think that for us was a big behaviour shift, even in how we kind of go to market. So it was a little bit around, as I mentioned before, having calls with clients, trying to build in the trust, social proof, and trying to understand what the individual or the business owner might be thinking in terms of what's stopping them from making that choice

[00:16:15] and how can we help to guide them through. And also, I think that's also from our perspective, for certain areas, we do have more humans in the loop. So for companies, there's a bit more human time, a bit more accountant time, reviews, calls, et cetera. But in the personal tax side, because we don't have that much of a behaviour change, that's a little bit more individuals just far themselves. So it's almost like each service that we provide now needs a tailor-made approach,

[00:16:45] which is based on the individuals and our clients and how they think. Now, as taxed will inevitably process more and more information and learn from more and more workflows, does data become a competitive advantage? And as you move ahead, how can you make the platform smarter over time while maintaining the privacy, security, and trust that customers expect? I would imagine it's quite a challenge to rein yourself in and innovate,

[00:17:15] but also get the balance right. Yeah, 100%. And I think for us, I think in the tax world, in the tax industry, the trust element, as you mentioned, is as important as creating new innovative tools and features for our clients and being able to use that data, handling it in the right way is super important. And aside from information security, like from a data perspective, it's important

[00:17:44] that we can use this data to actually help someone's tax filing, spot patterns around what trips a taxpayer are and also help around tax efficiency. And this is what we're currently doing in terms of adding in prompts, questions, utilizing aggregated, but anonymized data around what other landlords might be claiming back, what other sole traders might be doing to be able to provide insights on some complex cases

[00:18:11] to individuals. And I think that's where it's been able to turn it from just an individual's tax filing into something a lot more where they're getting specific insights on their case. It's as if they are working with a tax expert for their specific situation. And I'd love to leave startup founders that could be listening here a little bit of a takeaway. So if they're listening and they're looking at regulated industries, whether it be tax,

[00:18:41] finance, or insurance, and they see opportunities for AI-led disruption, what lessons from building tax would you encourage them to learn before they begin? I'm sure you've picked up a few lessons along the way, but anything you'd share? I think the key one from our perspective would be having experience in the industry and almost finding a niche in that industry. So every industry has their niches, whether you're in accounting, tax, insurance, et cetera, et cetera.

[00:19:11] Find that niche within your industry and try and create a beachhead of that niche. Make that the area that you own. I think for us, that's really served us well. In the expat world, we had that niche expat tax knowledge and were able to translate that into the B2C world and grow from there and now support various other industries as well. But that really helps us to gain a bit of a beachhead. I think that then speaks to the point

[00:19:40] around how in general we're always told to try and service the 80%. But actually, I guess if you service the 20%, those edge cases, those might be the ones who are willing to be your actual, what's the term, your early adopters, the people that are willing to try something new and to pay for it. So, could be worth exploring that kind of a thing depending on the industry and what you're starting. Excellent advice.

[00:20:10] I think that's a perfect moment to finish on. But for anyone listening that would like to learn more about tax, follow you online or just find out more information about anything we talked about today, where would you like me to point everyone listening? So our website is a great starting point, taxed.co.uk but also we're active on social media too so please do connect on LinkedIn and Instagram as well. We post a lot of content from a tax perspective so hey if you want to see some fun tax content we're on there too.

[00:20:40] Awesome. Well thank you so much for talking about your founder focus, the journey you've been on, the wider ecosystem for the future and some of the misconceptions around AI in tax and indeed the workplace. Really enjoyed chatting with you today. I'd love to stay in touch and see how things further evolve in this industry but thank you for sharing your story today. Awesome. Thank you Neil. Thanks for having us on and it was great to speak with you too. So many big takeaways from the conversation with my guest today. One being that useful AI

[00:21:10] begins with knowing precisely which job the system should perform. And tax sees a future where routine compliance can yes run with far less manual work but advice that still calls for context experience and someone that's willing to own the answer and I think that distinction matters well beyond tax when we're looking at AI in the workplace. And for founders my guest also offered a smart reminder the awkward edge cases

[00:21:39] can be a strong place to begin. And remember you can learn more about tax over their website. I'll include links to that their socials and indeed my guest LinkedIn. So a big thank you to him for joining me on AI at work and for making tax technology sound surprisingly and refreshingly human. But over to you where in your business could automation remove the repetitive work without removing the judgment that people and your teams rely on.

[00:22:09] Love to hear your thoughts. Techtalksnetwork.com go over there have a listen go over there there's lots of resources 4,000 interviews and also how you can meet me on the road as I take my podcast overseas at tech conferences work with me or send me an audio message whatever it is I'm here for you but that's it for today. Speak to you again real soon. Bye for now.