Tax technology turns over seemingly every couple of months, and, as a result, the role of the tax technologist is being transformed in real time. We wanted an inside perspective on these developments, so Michael Levin-Epstein, senior editor of Tax Executive, interviewed Nolan Ogden, PwC’s US tax technology leader, who has been with the firm for almost twenty years.
Michael Levin-Epstein: Let’s take a look at the role of the tax technologist. How do you think it’s changed in the last ten years?
Nolan Ogden: We see that our most valuable resource continues to be tougher and tougher to find and keep. Traditionally, we would take a CPA and train them to do multiple functions, whether it’s reporting, consulting, or data wrangling. With the dwindling [number] of CPAs and the increased demand tax departments are experiencing, we’re bifurcating certain functions and rethinking traditional CPA roles. We’re finding that some of these functions can be performed by talent with different skill sets such as engineering or data engineering. Not only are we seeing the shift with us, we are seeing this with our larger clients as well. Regulatory reporting requirements keep increasing, and the traditional tax talent pool continues to dwindle. The question has become, How do you reimagine these processes and functions with someone who isn’t a traditional tax expert or CPA?
Investment in Generative AI
Levin-Epstein: PwC is investing about $1 billion over the next three years in generative [artificial intelligence]. What do you hope to provide for your staff and your clients?
Ogden: The technology behind generative AI is not necessarily new, but the user interface placed on top of these models really brought the power of this technology into the spotlight. As a firm, we are building our reputation as a leader in artificial intelligence. Over the next three years, we’re investing a billion dollars in expanding and scaling our AI capabilities, not only to help our clients transform their businesses but our own as well. We call ourselves “client zero.”
We’re working through this journey ourselves, and everything we are learning we are applying to our clients. As part of this investment, we are going to upskill our entire firm through our My AI initiative, giving our people the knowledge and skills they need to work smarter and to help them grow their careers in different ways. From there, we can really help our clients understand the benefits of AI as well as other transformative technologies that are out there.
Levin-Epstein: What are the different ways that AI can help ease the burden on the tax function?
Ogden: One way is applying generative AI and its ability to generate tax content and process documents in ways we could not do before with machine-learning or other existing technologies. Generative AI models are trained on large amounts of data, including regulations, cases, and other content. A user can prompt the model and it will generate content, whereas before, a user would have to write that content and review. Now we have the potential for a machine to generate a first draft of content for a tax professional to interpret. For example, we upload a sale purchase agreement into the model and ask, “What indemnification does the buyer have, and is this a buyer-friendly document?” The model will analyze the document, summarize the indemnification sections, and provide a summary of the clauses and its predictions regarding friendliness to the buyer. Previously it would take an attorney or tax professional time to read the document, summarize it, and draft findings. Second, Pillar Two is on everyone’s mind right now, as it will fundamentally change the way we think about tax. Having a model that is trained on the rules and draft disclosures will help draft specific disclosures and so forth.
Enhancing Business Operations
Levin-Epstein: AI is just everywhere these days, and I am sure that there are ways that generative AI can help integrate and enhance other aspects of business operations.
Ogden: I agree the discussion is everywhere. I see that our clients are at different stages of maturity. Some are first movers and some are taking a wait-and-see approach. The most important thing is to keep tax aligned with the C-suite, because an overall technology transformation strategy is key. Bringing tax in, having your VP of tax sit in the same room while the C-suite is talking about the strategy and impact is important. One item to think through is if you are going to invest as a business in reimagining the front office, or even the back office, using AI, you might be able to generate R&D tax credits to help pay for this investment.
Levin-Epstein: Could you give us another example of a PwC AI success story?
Ogden: Sure. For perspective, generative AI is the rage at the moment, but at PwC we have been embedding AI into our tax tools for almost a decade. What we’re finding is that taking some of the traditional machine learning and then applying generative AI techniques can enhance the overall effectiveness. This media attention has actually put a broad focus on AI, which is great, and there are many tools to help our professionals work more effectively and get better-quality data and materials. Bringing that example to life, we worked with a client that has an online marketplace and is struggling with sales and use tax compliance and the taxability of specific transactions. The volume of transactions creates a classification nightmare. We developed a machine-learning algorithm to predict the taxability of these transactions, which worked fairly well. When we applied generative AI to the traditional model, the accuracy was greatly enhanced. This is a very specific example of solving a tax problem with traditional AI, but layering on some of these new techniques and making the solution more effective.
Levin-Epstein: There are a lot of skeptics who worry about reliability and accuracy when it comes to AI. What would you say to those skeptics to convince them that AI really is, for the most part, reliable and accurate?
Ogden: I am still not convinced that the technology is 100% predictable and reliable. I am, however, convinced that technology could give us a head start on solving problems. As a firm, we’ve had the most success pairing tax professionals with our data science team. Together they run prompts, interpret the results, and refine prompts until they are satisfied with the results. To me, AI is going to do great things on its own and has the ability to be revolutionary when paired with our tax professionals.
Levin-Epstein: Thank you.