How artificial intelligence will transform tax | EY
Susannah Streeter
Hello and welcome to the EY Tax and Law in Focus podcast. I’m Streeter, and in this edition, we’re going to be finding out how artificial intelligence, and specifically generative AI, has the potential to revolutionize the world of techs, perhaps even beyond our wildest imaginations. There are already big changes taking place. So-called GenAI copilots are already in use, sitting beside the tax practitioner, enabling non-IT specialists to analyze and manipulate core data sets, generating powerful insights. But there is so much more potential. EY research shows that the typical tax team spend somewhere between 40% and 70% of their total time gathering and manipulating data. Tax teams face increasing levels of complexity around regulation and real-time reporting, downward pressure on budgets, and increased costs of keeping tech updated. Then there’s the ongoing headache of the growing skills gap, an insufficient number of practitioners with the experience to meet the demands of tax authorities. We’re going to explore how, in the near future, GenAI will be able to help solve many of these challenges. I’m really delighted to say we have a panel of subject matter experts who are uniquely placed to offer insights into the opportunities which are there to be seized and why they could make such a difference to the tax function.
Streeter
But before I introduce them, please remember, conversations during this podcast should not be relied on as accounting, legal, investment, nor other professional advice. Listeners must, of course, consult their own advisors. Now, we have so much to talk about. We’re going to split this subject into two episodes. In round one, I’m very delighted to welcome Ricker, EY Global Vice Chair, Tax. Hello, Marna. Where are you today?
Marna Ricker
I am in Minneapolis and happy to be here.
Streeter
Fantastic. Great to have you with us. Also, welcome to Garcia Cure, who is EY Global AI Tax Leader. Hello there, Catalina. Where are you joining us from?
Catalina Garcia Cure
Hi, everyone. I’m joining from the coffee region in Colombia.
Streeter
Fantastic. Great to have you with us. And welcome to Duncan, Head of Thought Leadership on the Future of Work from Microsoft. Where in the world are you, Matthew?
Matthew Duncan
Hello, everyone. I’m actually out here in Seattle, Washington.
Streeter
Great to have you on this podcast. So much to talk about. And finally, a warm welcome to Campbell, EY Americas Tax Technology and Transformation Leader. Where are you speaking to us today from?
Daren Campbell
I’m joining us from Washington, DC.
Streeter
A warm welcome to all of you. We’re going to cover so many different subjects, but let’s look at the current state of play with regards to why AI could be such a game changer for tax leaders. Marna, let me start with you. How would you say GenAI is changing, challenging, and transforming the tax world?
Ricker
Thank you, Susannah. It’s really having an impact on every single thing we’re doing. Whether that’s tax authorities or corporations, it’s all about how we record tax, how we report tax, how we pay tax, and how we collect tax. It’s around data; it’s how we support and advise our businesses and the decisions they’re ultimately making. There’s an enormous amount of pressure out there. You said that in your introduction, from global tax policy changes to increasing tax collections and controversy and talent shortages. We’re going to get into that conversation and how GenAI is really changing and helping there. There’s economic uncertainty, geopolitical uncertainty. There’s a lot going on for tax departments, and GenAI is making a really big difference around that. As you’re getting into this topic, there’s a tendency to look at really short-term efficiency gains in my counsel. We’ll all talk about this as you have to look at long-term transformation. What really can you do within your department, within your people, to get after a long-term impact?
Streeter
Absolutely. We’re going to be looking at those long-term gains as well. But what would you say the value that you’re already seeing from GenAI for clients, as well as the business you lead at EY?
Ricker
There are three big things. So the first is just the everyday tools. Like, what are we doing every single day? So think desktops and platforms and technology tools and hands-on keyboards. And Microsoft has amazing things, things we’re already using in copilots, helpers built right into those desktops and the tools that we’re using. And think tax advice. And so really amplifying the technical expertise of our talent. Really amplifying, pumping that up, making it even better if it could get any better. Then, think really about that compliance and reporting piece. There’s a lot of data that we use in that process. Think about those process becoming really, really efficient and really quick and really accurate. Pushing the boundaries around those pieces. Those are the big three I’d want you thinking about when we think about what it’s already doing for our business anyway.
Streeter
Marna, thank you very much for the moment. Let me bring in Matthew. Matthew, do you at Microsoft have a feel for those industries and indeed departments or functions within organizations which are early adopters and those which are laggards?
Duncan
Well, we’re seeing adoption growing across all industries, specifically those with a large percentage of information workers. It’s where GenAI becomes that personal assistant that truly augments and takes away the mundane tasks and accelerates productivity, such as consultation or analysis. We’re finding that legal and tax professionals are, in many regards, leaders in this AI transformation. There’s definitely a mindset, those organizations that look at this as a true opportunity to reinvent and rethink the way they’re doing their business versus those that are standing back on their heels a bit and just questioning, is this just another technology trend?
Streeter
Thanks, Matthew. Catalina, to what extent do you think tax leaders think GenAI can provide solutions? Are they underestimating the possibilities?
Garcia Cure
A hundred percent. There’s a lot of underestimation. We’re seeing that the level of awareness and interest of tax leaders in the potential of generative AI is definitely increasing. There’s a very clear tendency to underestimate the full scope of its capabilities. Some people think that this is due to the natural caution that surrounds our field, but I think that it is primarily due to a combination of two factors. The first one is its lack of exposure to those applications that are truly, truly transformative. The second one, which we see every day with our clients, our role is to really clarify this point, which is that Generative AI is not just a more advanced version of classical machine learning. One very common reaction that we get from our clients is, yes, but we already tried to do that with AI, and we’ve been working on that. But no, let’s say that generative AI is a subset of AI, but what it can do is fundamentally different, almost as if you were talking from the user end or the company as a completely different technology in terms of what it can achieve.
Garcia Cure
We are not used to working with something that can process our language and mimic our creativity. Across the globe, we’re working with clients in use case discovery to help them understand the differences. So, machine learning in the tax function followed specific rules for finding patterns and predicting based on past data. But Generative AI can now, by itself, understand the reason and make decisions across a very broad range of complex tax activity. I’ll give you an example. It’s a real example from a client, but it also helps us bring awareness to other clients. If you tell them you have a process, and you’re manually and in a very slow way, so two weeks more or less, mapping provision data from multiple jurisdictions to a model that was taking two weeks and the effort of 6-8 people, as I just mentioned, and now it takes minutes. Our role is really to help our clients understand these possibilities. When they see it, they cannot unsee it.
Streeter
Amazing. So, Daren, Catalina is outlining there that leaders are underestimating the possibilities, and there is such potential out there. From your experience, do you think tax leaders are preparing their teams sufficiently to leverage the power of GenAI, given that so many are underestimating it?
Campbell
I would definitely agree that many of the leaders are underestimating the power of AI. We are seeing many companies that are on the journey, but most are just getting started. They’re just in the early stages. We see that in two forms. One is around training. They’re providing basic training on what it is, trying to help their teams understand what it is and some of the art of the possible around it. One of the areas that they’re focused on with the training is around prompt engineering, really how to communicate with the AI. This is a different way that we’re interacting with the technology than we’ve ever interacted before. That is one of the areas that we’re seeing a lot of focus on, and I would agree with that focus. I think that’s going to be a critical skill set. How we interact with the AI is going to be really, really important. I agree with that focus. The other thing that we see, as Catalina talked about, is looking for initial use cases. Tax is really ripe. There’s a lot of low-hanging fruit that can be picked up with AI. They’re beginning to look at how we still continue with our existing processes, but we find a point problem in bringing AI to that solution.
Campbell
We don’t yet see that many that have moved on in the journey where they’re really looking at the transformative power of AI. There are a few, but one of the things is that as companies progress and go through the cycle of basic awareness and beginning to identify point solutions, the next stage is really looking at how this becomes more transformative. One of the gaps that we have in tax is around data. We do have a few companies that are beginning to realize that and really beginning to look at the data life cycle, beginning to look at their data sources and how they’re managing data as an asset within the organization. Then, looking at this also being transformative, looking at changing their operating models. I think one of the things that’s really key as we talk about the transformative nature of AI is really a shift. This isn’t just implementing technology into the existing processes. This is really around rethinking those processes. There’s a framework that I like around this. What we’re doing right now is we have people who are executing processes that are presented with data that’s powered by technology.
Campbell
The AI shift is technology will be powered by data, and it’s executing those processes that are managed by people. I think that’s really key. There is a fundamental shift that we’re undertaking of moving from people executing the processes to really the technology executing the process with the people managing those. That’s a transformative shift. If we look at where people are today, we’re on the very early side of that. We haven’t really seen that, but that will be a critical step as we move to AI transforming the way that we do business.
Streeter
Okay, so Marna, Daren’s talking there about transformative power and rethinking processes. Let’s dive down into what concrete changes we could see due to GenAI, not just for the tax function but also how you could influence operations across broader organizations.
Ricker
I love that, and I get to work with Daren every day, which is one of the best parts of my job. One of the ambitions, certainly, that we have, and we’re seeing it already inside our own organization, but certainly one of the things that we’re seeing with some of the cutting edge departments that Daren’s talking about is the ability to take that data. Look, remember, tax deals with the lowest level transactional data within an organization. Because that’s where you have to report tax. Think about a VAT tax or invoicing. We have to report tax at source of a transaction for a company. That data is rich with insights for an organization. And the cutting-edge tax departments know that. And so they’re really, really looking at ways to provide good business insights right back to the company. And so you see things about the business inside that data. And so the ones that are doing exactly what Daren was talking about, letting technology lead the data process, letting humans manage that process, use the judgment piece of that, as opposed to doing that actual work hands-on keyboards. And so the tax departments can do that and then use that data to provide insights back to a business unit about their product or about something they’re seeing in the data itself, I think, are going to provide extraordinary value back to the organization.
Ricker
That’s really what we’re starting to see in the heading edge departments. It’s also what we’re doing here at EY. Again, tax is becoming the epicenter of data within our own organization that we’re able to then use to provide value back to all of our different service lines in the entire organization. That’s the power. That AI has the ability itself to summarize vast amounts of information, to extract extraordinary pivot tables and details and extensive documentation and data sets and efficient comparisons and predictive about what’s going to come next or what it’s seen in the past. And so that’s what we’re getting in these refined insights and refined information. And that’s what’s really, really powerful. So it’s got some limitless possibilities. I don’t even want to predict. I’ll let Matthew do a little bit of that, predicting what might come next for us. But I’m really trying not to limit its power within our own organization.
Streeter
Well, let me bring in Matthew. What’s your take on the potential for this limitless possibility? Do you think companies are underestimating the level of value that could be created and how overhead costs could be limited?
Duncan
Yes, I’m very much aligned with exactly what Daren and Marna had just spoken about. We see this role of AI in three areas. One, it’s definitely on this personal AI. What does it mean for the end individual, and how does this augment and truly make their work or amplify their expertise from going from good to exceptional? Then we see this foundational or functional AI happening at this department level where it’s integrated into the patterns of work, how we work as teams and within those organizations, and then a broader-based organizational AI, which is exactly what we are hitting on is how does it change the fundamental processes in which we actually get outcomes in the organization. But where we are today is that we’ve spent the last several months at Microsoft really understanding where users and companies are. We have the Work Trend Index, which is our survey that goes up to 31,000 people across 31 countries. And what we found was that 70% said they were more productive with Microsoft Copilot, and 68% said Copilot improved the quality of their work. So it’s just not the productivity, but it’s actually the enhancement of the quality.
Duncan
And we think that really, on a personal level, has occurred because it’s alleviating some of the challenges that we have in our work lives today. It’s 64% spent less time processing email, which is a real challenge; 71% saved time on mundane tasks, and ultimately, 75% spent less time searching for information in those files. We were at a point in which we almost have this digital debt. Generative AI breaks through that noise that we have in our daily work cycles, finding the information that we need, getting to the answers, creating and refining the work that we’re doing, and even in meetings. So you can imagine how that transforms us into more productive and, frankly, satisfied workers. But the real opportunity, going back to the points made, is not just in personal productivity, but to lift the capabilities of the entire organization. That’s where we’re seeing, really, organizations start to move forward now.
Streeter
Do you think that there is a lack of understanding about the tools available to leverage the power of AI and breakthrough that digital debt you’ve just been talking about?
Duncan
Yes. Back to a point that Catalina made earlier. This is not just another tool. This is a new technology powered by the force around natural language. I mean, the ability for anyone to use their natural way of communicating to get the answers and the reasoning engine that falls behind. Those organizations that think of AI as just a tool will fall short because it’s really a new way of working, and it’s going to take a mind shift and it’s going to take the creation of new work habits. I tell our customers to think about Copilot like a chief of staff. You need to provide context on what you need, and it just doesn’t end with one single brief prompt. You need to have a conversation, going back and forth, diving in deeper, and getting to the real outcome that you’re looking for. You can imagine how this true enhancement is going to democratize everyone because this can be an intelligent assistant for everyone in every department across the entire organization, lifting everyone up. I do think there truly needs to be a better understanding of that potential and how it’s going to reimagine and reinvent the way we’re going to run our businesses.
Streeter
So Daren, Matthew there is talking about our future working habits and the new chief of staff arriving next to us to copilot with us every single day. But what about current working habits? I mean, what does EY research say about the time that the typical tax team spends gathering and manipulating data, which could be better spent on more strategic activities?
Campbell
There’s increasing pressure for the tax function to spend time on more value-add activities. To your question, the surveys that we’ve done suggest that tax functions spend up to 70% of their time dealing with data. It’s gathering it, it’s transforming it, it’s analyzing it, trying to work through that data. When we survey some of those same companies and say, What type of shift of your time would you like to see in more value-add strategic activities? The initial shift is about 10%. They want to reduce the non-value add activities by about 10% and increase the strategic time and focus by the same amount on the other side. When we look at just the nature of the activities and the types of activities that tax performs, today, the split is about 72% of the time spent on one of those routine compliance-type activities, with only 20% on the high-value activities. And that 72% takes a lot of time. So, as we talk with companies, one of the challenges that they’re struggling with is the amount of time that it takes to prepare returns and go through the other compliance activities. It is so time-consuming that it does feel a little bit daunting to them about, Well, how do I upskill?
Campbell
How do I find time to shift and think more strategically? As Matthew talked about, that’s one of the beautiful things that we’re seeing about this technology is they don’t need to learn a programming language. This is skills that they’ve been using to interact and work with their staff and coworkers all along. Just now they were able to interact with this new technology using those same skills.
Streeter
Catalina, given the challenges that Daren has outlined there, it is clear that many of us would really enjoy having a wingman or wingwoman to help us in our day-to-day tasks. What examples are there of teams who are already making significant progress in exploiting this technology?
Garcia Cure
Sure. There’s too many lenses through which we can understand really the potential of generative AI for techs and provide some examples. I would use the categories that Matthew referred to a moment ago, which are personal and foundational AI. The first one, you said it, this personal assistant. We’re talking here about personal AI. As a personal productivity assistant, I think, in general, I’m maintaining the talent of our people in the form of a bot or a layer that’s embedded in the technology we already use. But think about a tax agent that also has access to general knowledge that’s always there to help you research, retrieve, analyze, draft, and translate faster and with better outcomes. My point regarding your examples here is that every single task can be a great example. That is huge. Everything that our tax professionals and the tax professionals at our clients’ tax functions do can be augmented with AI. If you think about it, not only is it amazing in terms of what the tax function can do, but it is also a privilege that not every function or industry could have. We are knowledge workers, and this technology is really powerful for workers who are working in their day-to-day lives with rules, language, and data.
Garcia Cure
As a personal productivity tool, everything is an example. I would say that the second one, as Matthew said, is the foundational AI, Generative AI as an embedded agent to automate and open not people but processes. Ultimately, it is people, but more holistically, it is a complete process. What we’re seeing at clients is a deep interest in integrating this technology within tax systems, going far beyond what we used to do before, which is simple rule-based automation. It’s not only about doing things faster, doing things that took more time before. There’s a lot of that, of course, savings in terms of time and funds, but also, it’s doing things in a new way and in a way that we couldn’t do before. Our teams are working on a few examples in which you can really grasp the full potential of the technology. You’re seeing generative AI from multiple perspectives, serving as a tool that adapts to many, I would say, challenges of the tax function. For instance, building a tax AI agent that can continuously monitor tax regulation changes across jurisdictions. Another agent can interact with humans and notify them of those changes and perhaps even recommend the next best action based on that regulatory change.
Garcia Cure
You could even go a step further and have a separate agent automating and updating systems and processes to ensure real-time compliance. We’re talking about a process here and really augmenting that process with the help of AI.
Streeter
A few innovations there that Catalina has outlined. Daren, what other developments are you seeing that have really captured your attention?
Campbell
Marna, I love the comment that you made about tax being the epicenter of data in an organization. That is true. Tax is one of the heaviest users of data in an organization. One of the biggest challenges that tax faces today is that we don’t own or control the upstream systems. As Catalina was talking about in her examples about getting to real-time compliance and being able to understand and have agents that are monitoring some of the regulatory changes in more of a near-time, real-time, this is one of the opportunities and seeds of innovation that we see in the tax function. Tax, as we go through the data, we’ve often been very reactionary. We’re looking backwards at transactions and data in the past, and then we’re trying to determine and then the tax treatment of what happened in the past. The opportunity for value that I was talking about before is really getting to when we get more near-time, real-time information; that’s when tax can really be strategic across the organization, not just related to compliance. There are a lot of things in tax where we’re looking at the nature of the transaction. But again, typically, we’re looking very downstream in the system.
Campbell
And that’s because of the amount of time it takes to review and process all the records. One of the areas of innovation that we’re seeing is in the accounts payable process. That initially upfront, when invoices initially come into an organization, on that invoice, there is transactional information that serves a variety of purposes downstream for tax. We have things like sales and use tax and VAT. We have fixed assets and how they’re coded, also related to property taxes as well. With the use of AI, we’re able to process and look at that information upstream in the accounts payable process to actually do some of that tagging and classification in more of a near-time, real-time area. That will enable or allow tax and the value of tax that it brings to be used in planning activities that today we’re generally not able to look at. That’s why we have a lot of work that’s spent today looking at refunds and filing amended returns; we’re catching things in the past. As we move that upstream in the system, we’re able to get better information, we’re able to keep more cash in the door for the company and to help them do a better job paying the appropriate tax for the major restrictions that they’re operating in.
Streeter
Great. Thanks, Daren. Matthew, both Daren and Catalina have talked about how GenAI is changing the day in the life of professionals. It’s clearthere will be productivity gains, but should we be looking a lot wider in terms of our AI calculations. Daren, there making that example of how you can delve back into the past to find efficiencies as well.
Duncan
Absolutely. Back to what Daren just pointed out, it really is AI is becoming interwoven into every aspect of the firm, but it’s reinventing the core business process where we’re going to see the massive gains, the massive value created. Let’s take other examples like global contract management. I know inside Microsoft, with 300,000 plus employees, there are so many people that touch the contract process. That puts a large load on our centralized legal organization to really make that all happen. You can now put the power out to the people through utilizing technology, something we’re calling Copilot Studio, that allows you to create low code capabilities that land on the desktop of every employee who deals with contracts. You can search, verify, and do a lot of the steps that have traditionally needed to be managed by a centralized organization and put that in the power of those across the organization, creating massive efficiencies at scale. These are the types of breakthroughs that we’re seeing. It goes beyond just the individual productivity. It’s really around those processes. I think the power of giving those processes or access or an action step to individuals so that they can be more self-served and be more active in getting what they need to get their job done.
Duncan
This is going to be transformative.
Streeter
Marna, what other business processes could end up disappearing altogether because of GenAI?
Ricker
I don’t think it’s about what disappears altogether. It’s exactly how Matthew and Daren framed it. It is far more about lifting up the human so that technology leads the process and the human applies the judgment. What I’m most excited about is that complete reshaping of what the human gets to do. What we get to do will get completely reshaped. That’s what I’m most excited about. I don’t think it’s about what goes away. It’s really about what we get to do. I don’t think there’s anybody who enjoys really, really manual, mundane things over and over again. I think that’s what’s exciting about it for me.
Streeter
But, Matthew, will this eliminate headcount then? I mean, those mundane tasks are currently carried out by people in jobs.
Duncan
Well, Marna hit the point exactly. It’s going to actually reshape organizations. We see a more dynamic, more global organization where it’s going to take and find efficiencies in one part and then redeploy talent in new ways to benefit business in other parts. Sure, there might be reductions in one area and additions in another. We see that in every technology inflection point. But we really see a reshaping happening, and that gets us really excited.
Streeter
So, the great reshaping of organizations is underway. Clearly, as you say, Matthew, there is plenty to get excited about. In the second part of the podcast, we’re going to focus on some of the challenges that GenAI could help solve. For now, Matthew, Marna, Catalina, and Daren, many thanks. And can I just remind you that the views of third parties set out in this podcast are not necessarily the views of the global EY organization nor its member firms. Moreover, they should be seen in the context of the time in which they were made. I’m Streeter. I hope you’ll join me for the next part of the Tax and Law in Focus podcast, focusing on how GenAI has the potential to revolutionize the world of tax. EY, Building a better working world.
