The inaugural PwC Global Reframing Tax survey, which gathered perspectives from more than 1,200 executives across 47 countries, has uncovered a high rate of artificial intelligence (AI) adoption by tax functions across the globe. Some 56 per cent of tax professionals said they were already seeing tangible benefits from GenAI, while more than 80 per cent expect it to transform tax planning and strategy in the next three years. However, less than half, 47 per cent, saw leveraging new technology in the tax function as a top priority.
“We’ve been talking to a lot of companies in Ireland about how they are using AI, particularly in the tax function,” says Harry Harrison, tax partner with PwC Ireland. “Three issues are coming through. Tax functions are facing demands for greater efficiency. They are dealing with massive regulatory changes while at the same time there is a push for the tax function to play a more strategic role in overall business transformation.”
Fortunately, AI can help with all of those issues, he points out. “It can sort and analyse data very quickly, it can analyse regulations more efficiently, and that frees up time to allow the tax function to take a step back and think more strategically.”
Harrison believes the low priority being placed on AI adoption by some tax professionals is to be expected at this early stage in the technology’s evolution. “Tax professionals are accustomed to getting perfectly correct outputs from the technology they use,” he explains. “But AI is not yet getting it right all of the time. There is still a lot of getting used to the technology going on, and people are still getting their heads around that. The issue will be sorted over time, but people are naturally cautious and have some reservations about it at the moment.”
That makes the people part of the equation even more important, according to PwC Ireland director & Tax GenAI lead Karen Kiernan. “It’s not just about the technology,” she says. “It’s about unlocking the potential of the technology. That’s where people come in to understand the technology, make judgments understand the limitations. Tax functions have been using specific-task focused technology for a long time, but GenAI, in its current incarnation, is more amorphous in its application. People need to learn how to use it and how to work alongside it. That all takes time. Organisations need to invest in training their people for that.”
It is clear from the research findings that organisations understand that need. “The survey showed that the number one skill for tax professionals to acquire is AI - this is the main skill that tax functions around the world are now looking for,” says Harrison.
PwC Ireland made a conscious decision to become an early adopter of AI and GenAI so it could bring its own learnings to bear for its clients. “PwC has been using GenAI for nearly two years and the tax practice has been at the forefront of that,” says Kiernan. “Its functionality is at the sweet spot of what we do in tax. It can ingest huge amounts of data, analyse and summarise it, and we can apply that to client cases.”
The tax practice made the brave decision to invest in GenAI assistants very early on, she continues. “We trialled them with 100 staff, about one eighth of our tax practice. We could see the potential, but the technology was not quite there yet. We learned by trialling, but we needed a more bespoke GenAI to get better results.”
A eureka moment
That led to PwC developing its own bespoke tax research GenAI model. “That proved to be a eureka moment,” Kiernan adds. “It helps people in what they do day-to-day. It enables them to put queries and get answers. It’s opened people’s eyes to the power of AI, and they are now more open to using more general AI assistants. We have now made GenAI available to everybody in the tax practice and the bespoke tax research assistant is also available to everyone that wants it.”
The response has been very positive. “There is naturally a fear factor around this technology,” Kiernan accepts. “But our people’s experience with AI means they are starting to trust it more. It’s been hard work getting there in terms of demystifying the technology and taking the fear factor out. It is changing people’s roles, but it is not replacing them. How they are performing tasks has changed. They are spending less time on data processing and moving up to more value-adding assignments. They are getting to do higher-value work earlier in their careers and they are enjoying it. People become more valuable to the team and have more of a voice in the team. It’s giving people a better working experience and we are seeing usage increase significantly.”
The adoption journey hasn’t been without challenges, however. “It’s not been without bumps in the road,” Harrison notes. “You have to manage people’s expectations. There was an expectation that you press a button and get output. But it has to be checked, and you will always need human critical thinking. The output is not always perfect, and it does have the capacity to hallucinate, and it doesn’t yet do nuance well. We had to explain and experience what it is and what it isn’t.”
He also explains that the change management process required is not “one and done”. “The technology is reinventing itself every six months and the change management process needs to be constant in response to that.”
No off-the-shelf solution
Those experiences will prove valuable for clients, he believes. “One of the reasons we pushed so hard for adoption is that as we’ve done the hard work and have seen the benefits and drawbacks in our business, we can have much more meaningful discussions with clients and how they can best utilise AI. The AI answer will be different for every client and the conversations we have with each client will also be different. There is no off-the-shelf solution. A company needs to identify their pain points, what they need fixed, what will make life easier for them, and then look at how can AI help on that journey.”
Kiernan agrees, pointing out that identifying use cases for the technology is crucial. “There has been a lot of hype, but that hasn’t really been matched by adoption. However, if you diagnose the use case and the pain points to be addressed and bring the right AI solution, that makes adoption much easier and more impactful.”
She also stresses the importance of data management and governance. “Data is the fuel for GenAI,” she points out. “You really need to invest in a data management and governance strategy so that you can give properly organised data to fuel GenAI and get the best out of it. Also, people using GenAI need to understand data regulations. There is a lot of training to be done there.
“We are only at the tip of the iceberg of what AI can do for tax functions,” she adds. “AI will revolutionise compliance and reporting activities, for example.”
For companies starting out on their AI adoption journeys, Harrison notes that the PwC Ireland specialist tax technology team advises companies to take a three-step approach when considering how technology such as AI can be used to maximise the efficiency of the tax function.
“The first is to identify the critical processes and the pain points you want to address,” he says. “Next is to look at the skills, people, technology and data you already have in place and how they can be leveraged to meet the challenge – that will help identify the technologies and tools you might need to acquire. Finally, build a clearly defined technology roadmap that addresses the critical processes to enable you to invest in a scalable solution covering all the information flows in the tax function.
“Organisations need to start the journey now if they don’t want to be left behind,” he continues. “By adopting AI tax functions, they can become more strategic and lead out on AI and help the rest of the business adopt it.”
Adds Kiernan: “We are optimistic about the opportunities AI can bring. The tax function is in the sweet spot for what AI can do. Key learnings for us are to manage expectations and be transparent with people to understand that AI will not replace them; it will enable them to do their jobs better and make their roles more interesting – and that can be very powerful.”
