CFOs are facing a tougher remit than ever. To start with, globalisation trends mean many large businesses operate across borders and are subject to an ever-expanding roll call of regulations, such as country-by-country tax reporting requirements and the OECD’s Base Erosion and Profit Shifting rules.
“That’s creating challenges, because the number of data points that companies have to collect is increasing at an exponential rate,” says EY Americas’ tax technology and transformation leader, Daren Campbell.
Tax administrations around the world are also increasingly digitising their operations, which gives authorities more visibility into a company’s finances while also speeding up the entire tax process. Take Mexico, for example. Its tax collections increased by more than a third when introducing its e-invoicing system in 2011. Mexican companies also have to respond to requests from Mexico’s tax administration service (known as the SAT) in a much shorter timeframe – about 10 days compared to 90 days in the US.
In addition to those pressures, the accounting profession in general is suffering from a talent shortage, which means tax and finance functions are typically under-resourced, making it harder to keep up with this growing workload.
“There are more data points that we’ve had to process than ever before, we’re having to do it at a faster pace than we’ve had to historically, and there are fewer resources available to do the work,” Campbell says.
In the past, companies have dealt with resource pressures by outsourcing to cheaper offshore service providers. As the amount of data finance teams now have to deal with is growing at an exponential rate, that strategy is no longer a sustainable solution. But AI might have the answer.
AI to the rescue
“Offshoring is still a linear approach, and you can’t solve an exponential problem with a linear solution,” says Campbell. “This is what AI really solves – it helps us get through greater volumes of data faster with fewer people.”
For example, one area where AI can improve efficiencies and reduce costs is by automatically categorising data and removing low-level data entry work that is slow to complete and prone to human error. By handing this job over to AI, something that once may have taken a human team 500 hours a month to complete can be cut down to a matter of seconds, Campbell says. In addition to more efficiently categorising data, AI can also reduce review times by giving a confidence score on how accurate it believes that categorisation is, enabling reviewers to focus on those uncertain cases and making the process far more streamlined.
The rapid advance of generative AI tools is also creating additional opportunities to improve the way tax and finance teams work. For instance, it can change how accounting professionals access knowledge, making it easier to search for information they need on certain tax laws or policies.
“The way we manage critical knowledge assets has always been human expert based with minimal technology enablement across increasing daily volumes of technical alerts and regulatory change,” says Richard Clough, chief data officer for EY Global Tax. “With GenAI, we can build a user-friendly prompt experience around knowledge access and provide a more scalable and robust approach to scale up the work of our tax professionals, made possible by greater technology enablement.”
GenAI tools can also help tax and finance teams better manage regulatory changes by horizon scanning for new rules and flagging critical areas, effectively shortlisting potential issues for a tax expert to look into more closely.
“Experts are effectively supported at scale, which can really expand the amount of quality work any individual can do, and therefore the impact of an individual gets magnified as a result,” says Clough.
Stages of innovation
Campbell says there are typically three stages to a company’s AI journey. The first level often involves AI assistant tools, such as Microsoft’s co-pilot or other chat bot-like interfaces, where users type in prompts and receive answers or recommendations. The second level is embedding AI into a discrete process where human intervention is reduced but the process largely stays the same. And the third level is where companies embrace an AI-ready data ecosystem that completely changes how processes are executed.
“Level three is really where AI becomes transformative,” Campbell says. “At level one and two, companies will get some benefit from that but it won’t lead to a transformation of the tax function or accounting function. Level one and two are really just stepping stones between each other. Between levels two and three, there is more of a chasm, and the bridge that connects that chasm is having AI-ready data.”
There are several challenges CFOs face to advance to that third level and start generating AI-ready data. The first hurdle CFOs typically have to clear is around governance and internal AI policies regarding what can and can’t be used. In most cases, this means CFOs are unable to just start experimenting with AI tools because internal rules are often yet to be fully agreed and established. The knock-on effect of that is most companies find themselves stuck in stage one where they are using digital assistants such as ChatGPT or tools they have built in-house.
A second key challenge is around broader change management. Campbell likens this to the Second Industrial Revolution when factories moved from steam-powered transaxle-driven machines to electric motors. In the first instance, the machines were still all hooked up to the transaxle. Then, factory owners realised by putting an electric motor in each machine, they could change how the factory floor was set up and even the way factories themselves were built, transforming the production process entirely.
“When we’re looking at the adoption of AI in accounting and tax, our processes have been set up to be operated by people,” says Campbell. “So it’s people that are doing the execution and technology is the tool. The shift with AI is that we’ll have technology, fuelled by data, executing those processes. And then, those processes will be managed by people.”
The data dilemma
The third and possibly biggest challenge to achieving level three that CFOs face is around data, which in many cases is a long way from being AI-ready. This is because data is frequently stored in different systems or trapped in spreadsheets sitting on an individual’s computer and out of reach of any AI tool. In addition, tax teams are often downstream receivers of that data, and those upstream systems are not designed for financial reporting, which means the data usually needs to be treated before it can be used.
“Maybe all companies are somewhere on the journey, but there are not a lot of companies that are very far down the road yet in having data that is accessible, cleansed and reliable,” says Campbell.
To advance down that road, companies need to develop a comprehensive data strategy. The first step in that process is separating data from technology and putting in place a data roadmap that outlines where they are now and clearly defines the desired end goal. This means understanding the entire data lifecycle, including who owns and is responsible for the data, and who consumes it.
“That needs to be the starting point, and most organisations don’t have a plan or strategy around their data,” Campbell says.
Taking advantage of AI-driven data agent tools can also help with large-scale data mapping and cleansing, as it can automate much of that process, Clough adds.
EY is also helping organisations navigate this process through its EY Data Labs unit, which helps organisations improve their data management capabilities when seeking to transform their operations.
“It recognises that the best solutions are more of an ecosystem approach,” says Campbell. “It’s not just about helping clients improve their data, it’s also about understanding their data challenges and then focusing on key areas and working with the client on finding the right solutions, sometimes bringing in a technology vendor that may specialise in those areas.”
Because many CFOs are facing similar challenges, this ecosystem approach means innovative solutions for one company can act as an accelerator for other businesses, helping modernise the in-house tax function at scale in a more collaborative fashion.
Efforts to modernise the tax function and rethink existing processes inevitably mean the role of the tax professional will also change in the future.
“In the past tax professionals took on the role of an archaeologist,” says Campbell. “We were always digging through the dirt of past transactions and trying to find the bones or pieces of pottery to determine what the appropriate tax treatment should be. Tax was previously always after the fact; now it’s going to be more forward-looking.”
That means the skills tax professionals will need in the future are also likely to change. While tax domain expertise is still going to be important, the advance of GenAI means finance and tax teams will need to improve their judgement, so they can better assess the responses AI gives and quickly understand what can be relied on and what additional information is needed, says Campbell.
Therefore, by using AI to remove repetitive mundane tasks and by honing their judgement and decision-making skills, tax professionals can start to provide more strategic guidance and advice and help the tax function become a more valuable partner to the wider business.
Find out more at ey.com/TFO
