Can GenAI transform tax teams? It all hinges on humans
August 2024
The personal computer, the mobile phone, the internet. It’s rare that innovation becomes a barrier to progression, so why should Generative AI (GenAI) be any different? Rather than technology replacing people, this is about unlocking our potential and reimagining how humans and machines work in tandem.
The evolution of skills is part of our history and thanks to the explosion of interest and investment in GenAI, it will happen again. As the tax profession embraces the tech, it will transform the nature of tasks – potentially bringing major productivity wins – plus the expertise needed for teams to do their work.
As well as new capabilities, and a shift in how work is performed and valued, any large-scale adoption will require substantial organisational change. For tax leaders, there’s a lot to get to grips with.
A new era of automation
GenAI creates content based on input questions, whether that’s text, images, code, audio, or video. Behind user-friendly interfaces are large language models trained on vast amounts of data to understand, and mimic, human communication.
Much hyperbole surrounds GenAI, but it’s clear there’s real potential for this technology to help better serve businesses.
For tax teams, think compliance, HMRC enquiries and risk reviews, and filings in specific formats, such as Country by Country Reporting. With the right prompts, it can generate everything from briefings and memos to training materials and summaries of technical documents.
But GenAI deployments will only be successful if the talent model can adapt to this new vision of work. Recent research by the Deloitte AI Institute showed that 75% of business leaders surveyed expect to change their talent strategies within two years.1
“When humans and machines collaborate, they can achieve something neither could do independently.”
Fresh possibilities for our people
Successful GenAI adoption is about people and machines coming together. Automating routine tasks will allow teams to focus on higher value assignments, creating new opportunities. But AI-generated content must be validated by humans to maintain quality and control, especially in tax environments where there’s a need for absolute, auditable accuracy.
To make that partnership work, organisations will need to attract and build capabilities not usually associated with tax. In the short term that means increasing AI fluency, so it’s likely recruiters will target people with less traditional backgrounds, for instance in data management and analytics. The modern tax professional should understand the fundamentals of GenAI including prompting and, most importantly, its functionality and limitations. But beware, the rate of progress is also rapid. With the launch of GenAI capabilities into PCs, prompt-based AI may become the preserve of data scientists/developers/power users. For the common user, GenAI will ‘just happen’. Organisations will need to determine the extent and depth of capabilities required and where they should be concentrated.
While tech can mirror human expertise, it cannot replicate the emotional intelligence and communication that clients need. During Deloitte’s research, when leaders were asked which skills would increase in value in the era of GenAI, data analysis came out top. But critical thinking and problem solving – two very human capabilities – were second.2
AI has been used in tax for many years, largely for compliance. But GenAI could allow tax professionals to expand into the advisory space, with professionals becoming more efficient in processes such as tax planning and scenario simulations.
There may also be a shift in the ratio of entry-level roles to senior positions. Certainly, organisations embracing emerging technologies are more likely to attract tomorrow’s brightest minds.
“If a task requires effort to execute but is easy to validate, it has the hallmarks of a good GenAI use case.”3
Collaboration, communication, integration
The technology cannot replace human experts, so there must always be a person in the loop. It’s imperative, therefore, that tax teams learn to use AI as the future of work evolves.
Tax teams will need to move out of silos and liaise more closely with other areas of the organisation, including IT and Legal. An integrated approach will push collaboration skills higher up the wish list. It could also increase employee mobility as new, cross-functional opportunities open up.
Coaching, mentoring and on-the-job training, such as that offered by Deloitte’s AI Institute, can upskill existing teams, while secondments and internships are an effective way to bring in new expertise in the short term. And looking to the future, preparing students for a world with GenAI will be vital.
To make sure tech deployments remain people-focused, leaders must take their teams with them on the journey, understanding what employees want, like and need to do to keep them motivated, engaged and fulfilled.
“Redistributing activities can create new opportunities for employees and make them even more valuable to the organisation.”
How trusted advisers can help
Shortly after ChatGPT3.5 was launched publicly on November 30 2022, it had amassed 100 million monthly active users. ChatGPT4 followed in March 2023, then 4-o arrived in May 2024. With each release, there are significant advances in the model’s ability to understand language, and its capacity to reason.
To keep up with the speed of innovation, the sooner tax leaders grab the reins, the better.
While they won’t be expected to become AI experts, they will have to understand how to integrate it into their function, and the pros and cons of doing so.
As quickly the then benefits of GenAI have been identified, the risks, limitations and ethical considerations have also become apparent. The International Monetary Fund, for example, has raised concerns over whether AI could widen the gap between rich and poor nations.4
And when it comes to the data on which GenAI is trained, there are questions around ownership, bias, legality and quality. Large language models are inherently predictive, so there is a possibility of them generating incorrect or misleading responses, known as hallucinations, by perceiving patterns that don’t really exist.
While some of the limitations can be overcome by introducing domain-specific data, or building bespoke models for businesses, it’s crucial that leaders know the risks and what constitutes ‘trustworthy’ AI.
So how do you start the GenAI journey?
Being part of a community with other tax professionals addressing similar challenges will be beneficial and trusted advisers can bridge the knowledge gap.
The Deloitte AI Institute is supporting clients at various stages of their transformation. For example, experts have designed a Digital Artifact Generation/Validation method to determine whether an idea can be turned into a GenAI use case.
Also, published earlier this year, Deloitte’s AI guide for Tax Directors offers valuable insights to help leaders develop a framework for thinking strategically about GenAI in the short, medium and long term.
For the tax profession the potential is boundless, but there’s a lot of work to do to understand the possibilities and shape the vision. If it feels daunting, there’s an old consulting saying that’s worth remembering, “How do you eat an elephant? One bite at a time.” This is certainly true of GenAI.
To find out more about how Deloitte can guide your GenAI journey, visit our Tax & Legal Digital Nerve Centre.
________________________________________________________________________________________
References
