White paper

AI and taxation: What professionals need to know

Introduction

Artificial Intelligence (AI) and machine learning (ML) offer substantial cost-saving potential in nearly every sector of the global economy. The tax and accounting space is no exception, and introducing AI into the industry has created some exciting opportunities for auditors, accountants, and other tax professionals as they look to optimize their operations. Below, we will discuss some of the ways tax and accounting firms are already implementing AI. We will also discuss other benefits, challenges, and areas of interest concerning this emerging technology.

One question that consistently arises among professionals is whether technology will replace their jobs. It is a question that frequently creates resistance to innovations and the benefits they can provide. The truth is that technology is intended to streamline manual tasks so professionals can focus on more meaningful client work.

McKinsey & Company recently laid out a very compelling case for how AI-driven automation will transform a vast array of industries, even ones that have primarily been considered insulated from the technology. As pointed out in its recent paper, “The economic potential of generative AI: The next productivity frontier,” previous iterations of this tech were geared more toward data management tasks like data collection and processing. However, advancements in areas like generative AI, which we will cover in more detail later in the paper, have allowed “knowledge work” areas to enjoy many of the benefits of the technology.

To that end, areas of work like “collaboration” and “decision making” have seen huge jumps in automation potential, notes the report. By its estimation, McKinsey & Company identified a 34-percentage-point increase in the potential to “automate the application of expertise.” Similarly, the potential to “automate management and develop talent” jumped from 16% to 49% between 2017 and 2023. These figures represent tremendous promise for the tax and accounting space. 

Understanding the basics of AI in the tax industry

To better appreciate the direction AI technology is heading, it can be helpful to consider some key concepts driving its adoption. As such, let’s take a closer look at what exactly AI is and how it works, and then we can explore some of its derivatives and related technologies.

What is artificial intelligence?

Simply stated, AI refers to “the ability of machines to perform tasks that normally require human intelligence, such as reasoning, learning, decision-making, and problem-solving,” explains the Blockchain Council. These tasks can range from creating artwork to editing copy and — drawing an example from the tax and accounting profession — reducing manual labor with respect to preparing an individual’s tax returns or auditing a business of virtually any size.

What is generative AI?

In addition to AI proper, several subfields of the technology have recently emerged. One such field is “generative AI” (GenAI), which refers to a category of AI algorithms used to generate specific outputs considering the data the generative AI system has received and been “trained” on, according to information from the World Economic Forum.

“Unlike traditional AI systems that are designed to recognize patterns and make predictions, generative AI creates new content in the form of images, text, audio, and more,” notes the international body.

This technology can be applied to a wide array of products and concepts, and it has become a popular tool for both professionals and amateurs in a variety of industries. Generative AI pulls from “generative adversarial networks,” which are basically derived from a special type of deep-learning process, explains the World Economic Forum.

These consist of two interconnected neural networks: one to create new data and another “discriminator” network to evaluate that data. Together, these two networks feed off each other to incrementally improve outputs based on feedback from the discriminator network. This process continues until content “indistinguishable from real data” is generated.

What is machine learning?

According to the MIT Sloan School of Management, “Machine learning is a subfield of artificial intelligence, which is broadly defined as the capability of a machine to imitate intelligent human behavior. Machine learning is one way to use AI. It was defined in the 1950s by AI pioneer Arthur Samuel as ‘the field of study that gives computers the ability to learn without explicitly being programmed.”

Artificial intelligence and its subfields are reshaping nearly every sector of the economy. You would be hard pressed to find an industry that has not at least explored how people could leverage AI — both generative and otherwise — and machine learning.

Even before AI entered the conversation as the veritable business soup du jour it is now, the International Federation of Accountants has been pushing for machine learning in the accounting space. For auditors, machine learning can be a powerful tool in data sampling, as “entire ledgers” can be quickly analyzed.

“In a more advanced application, a set of transactions could be provided to an AI tool, and machine learning would identify the patterns in the transactions and be able to identify what ‘normal transactions’ look like. Using this method, it would then identify those exceptions that don’t match the norm as exceptions,” reads information from the trade group.

This information is invaluable for auditors dealing with large amounts of data, which is often the case.

The journey from theory to practice has been, by many metrics, a rapid one. In the tax and accounting space, AI adoption has rapidly picked up speed in recent years, and some degree of integration has already begun for many firms. Let’s briefly explore how that journey unfolded to get a better sense of where it is going in the near term.

A brief overview of AI and its history in the tax and accounting industry

There is certainly no shortage of organizations working toward developing functional, safe, and effective AI. OpenAI’s popular ChatGPT program represents one notable effort. ChatGPT is a language-based AI model that has gained a lot of attention in recent months for its rapid integration into both the financial world and the cultural zeitgeist.

Consider the evolution of ChatGPT as chronicled by the Urban Institute and Brookings Institution’s Tax Policy Center. In the first five days it existed, ChatGPT had already gained a million users, and that number climbed to 100 million in just two months — then to an astonishing 1.16 billion users in the three months following those.

However, the Urban Institute and Brookings Tax Center points out that the road to its current iteration was long, and there were plenty of speed bumps along the way. The earliest versions of the program were only able to answer a handful of questions included on law school entrance tests, and it struggled even worse when it came to tax and accounting questions.

Eventually, those earlier versions saw major improvements, and ChatGPT-4 ultimately went on to pass the bar exam in the 90th percentile, achieve a near-perfect score on the GRE Verbal component of the graduate school exam, pass nearly all 15 advanced placement tests it was given, and perform tax-based calculations when instructed to operate as TaxGPT.

Tax authorities in Greece and France have used AI to cross-check property tax registries and satellite photos of homes to find tax cheaters who don’t declare assets like swimming pools. And Johns Hopkins University computer scientists are creating ‘Shelter Check’ to enable Congress, the IRS, or courts to scan legislation or rulings for loopholes,” adds the Urban Institute and Brookings Tax Center.

The evolution of AI in the tax and accounting industry: How is it being used now? 

As noted in our recently published piece, “The impact of artificial intelligence on the tax and accounting profession,” AI as a tax and accounting tool has rapidly gained traction. Why? As stated in the piece: “This technology offers a range of advantages that CPAs can leverage to provide better services to their clients. Automation of mundane tasks, real-time insights, and personalized advice are just some of AI’s benefits to the industry.”

To that end, AI is quickly becoming a powerful companion for tax and accounting professionals. But the technology, at present, remains just that: a companion. The most effective use of the technology is to put it to work as a super-powered sidekick. It is not built to replace tax and accounting professionals, and its current iterations are not designed to replace human labor.

At its core, AI is built to transform the work people — like tax and accounting professionals — do each day, and its greatest value is to provide a problem-solving solutions provider that facilitates an efficient, streamlined workflow.

Tools to enable smart guidance

Much like many other professional services, tax preparation comes with inherent requisite work that, while often mundane, must still be done quickly and accurately. It is here that AI shows its true worth. By automating tasks that do not demand as much specialized expertise as high-end, value-added advisory services, for example, tax and accounting firms are afforded much greater flexibility when allocating time and resources toward profitability and growth.

In the next section, we will discuss some benefits of deploying AI in tax and accounting service delivery. Ultimately, all arrows point toward boosting operational efficiency so you can spend more time and energy managing and improving relationships.

The benefits of AI for tax professionals: Improving efficiency, accuracy, and relationships

AI’s integration into tax and accounting services has helped enhance and improve many internal processes, but few are more prominent than task automation, data analysis, and both internal and external communications. This can be accomplished, explains the Blockchain Council, by taking enormous volumes of data — both structured and unstructured — and massaging that data with things like sentiment analysis and predictive analytics.

As the information from the council explains, “AI can automate tasks that are repetitive, tedious, or prone to human error, such as data entry, invoice processing, reconciliation, tax preparation, payroll, auditing, and compliance. This can save time and cost for accountants and allow them to focus on more value-added activities such as strategic planning, advisory services, and business insights.”

According to the Blockchain Council, a survey conducted by Sage indicates that as many as 58% of accountants reported they expect AI to automate almost all data entry tasks in 2023. Another survey showed that 79% of CFOs expect AI will help bolster their firms’ analysis and planning capabilities.

Part of the appeal of AI also lies in its capacity to process massive data sets quickly. The breadth of these data sets might be quite disparate, too, as they may range from things like financial statements and reports to emails and even social media content. But all the data gets processed the same, adds the council. An AI companion can quickly consume these data sets using techniques like predictive analytics and anomaly detection.

In addition, AI can also help improve how tax and accounting professionals communicate with one another and their valued clients. Natural language reports and summaries can be generated from numerical data sets, and AI’s natural language generation can provide effective chatbot user interfaces derived directly from its data sources. This, notes the council, will also help ensure clients have access to quick, easy answers to questions related to their accounts.

Improving the efficiency of your operation at these fundamental levels will go a long way toward creating a leaner, more productive, and, ultimately, more profitable business. Consider the following tips for information about effectively implementing these AI-powered efficiencies.

How to get started with AI in tax: Best practices and implementation strategies

Rolling out a new AI tech stack will look like rolling out most new business initiatives. It will require a holistic effort that considers human resources, costs, training, and support — and the continued evaluation of its implementation will be needed. In general, it can be risky to rush the implementation, especially if the staff has not been brought up to speed on the purpose and strategy of the initiative. For that reason, it is advisable that managers considering an update to their tech stack begin by evaluating who it will impact, how it will impact them, and what support those individuals might need throughout the transition.

In the next section, we will evaluate some notable best practices for implementing new technologies. While there is no absolute right and wrong way to transition to an AI-driven environment, there are plenty of great strategies managers across sectors have already successfully implemented.

What can organizations looking to effectively use AI do for a smooth transition

When adding new AI technologies into your workflow, starting small and building the tech stack unit by unit until you have reached your implementation goal can be helpful. Like any change in the workplace, AI integration usually works best when scaled up at a manageable pace.

This implementation generally occurs in stages. For many shops, this looks like a period of “upskilling” employees, a “collaboration” among the various levels of the firm, and “augmentation” to reinforce new workflow norms. As managers consider the value of implementing new technologies, breaking the process down into these digestible steps can be helpful.  

During this evaluation period, firms and organizations considering their needs can benefit greatly from checking in with their personnel departments to gauge available upskilling resources. Like most workplace changes, knowing where the team stands and what avenues to explore to get them where they are going is helpful. 

Questions to consider include:

  • What do they want? 
  • What do they need? 
  • What is the best way to get them to buy into the new system? 
  • What level of training will be required to make this all work, and how much support will they need after implementation?

Tax Executive, the professional journal of the Tax Executive Institute, offers some insights into leveraging existing staff and offers some things to consider as you onboard team members.

“Technology implementation is most successful when tax team members have previous experience with automated tools, strong Excel skills, and the ability to adapt to new software and workflow processes. The team also needs the capacity to manage its day-to-day work while transitioning to the new platform,” according to information from Tax Executive. “And don’t skimp on training. Success depends on team members’ comfort, confidence, and competence when the new system goes live.”

Equally crucial to the training associated with the upskilling phase of implementation is the value of bringing everyone up to speed together in a meaningful way. Collaboration breeds trust and creates a safe-feeling environment for employees, which will make subsequent augmentations attached to the tech upgrades stronger and more effective, too.

The International Federation of Accountants reminds managers to be cognizant of the workplace culture shock that is sometimes associated with a transition to new technology. Sometimes, there might be resistance to the initiative, and sometimes, that resistance might come from more than one level of the firm. Communication is critical, and it is always helpful to lean on trusted voices in the organization who can offer honest feedback about how implementation is progressing. These same voices can frequently serve as champions within the firm to support the new technology and its benefits.

Upskilling support for tax pros is critical during transitory periods

For those looking to bolster their existing skill set, there are lots of great resources available for perusal. Thomson Reuters recently published a piece entitled “How to upskill accountants” that also points out many organizations have training options and seminars available to support their team members. Team members that engage in such activities can also enjoy a welcome boost to staff engagement, morale, and the opportunity to bring a fresh perspective to their work.

It is also worth noting that while it is important to continually engage in professional development to get the most effective use out of new technologies, those same technologies also offer training benefits themselves. Application programming interfaces, artificial intelligence, and automation offer a great deal of opportunity to accountants looking to take their skills to the next level.

Do your homework to ensure data privacy and security

Data security and privacy are among the most important considerations tax and accounting professionals will face as they begin this journey. For this reason, it is also one of the most important reasons to find a trusted solutions provider to ride with you on your AI integration journey.

One important thing to keep in mind as initial discussions about the technology progress is that staff may not immediately recognize how AI implementation will impact data storage procedures. For this reason, it can be extremely helpful to ensure they understand the importance of protecting client data and staying abreast of new workflows that will impact where and how data is stored.

Having conversations early and often about data security and client privacy can help prevent many issues that might crop up down the road. This is another reason why ensuring your teams are aware of any changes born from implementing a new AI-based tech stack is often a priority for individuals and organizations making this jump.

For more on the importance of data security, follow this roadmap to AI integration.

Building trust in AI: The importance of transparency, accountability, and explainability

The effective implementation of AI-based tax and accounting systems takes careful planning. As with any other tech upgrade, things can sometimes get a little bumpy during execution. There are some inherent risks associated with implementing AI; lawmakers, policy experts, and business executives have each weighed in regarding these topics.

One such example is the guidance offered by President Joe Biden’s administration in October of 2022 when the White House published a comprehensive report entitled “Blueprint for an AI bill of Rights: Making Automated Systems Work for the American People.” Among the many areas covered in the document was guidance on how to protect data and ensure that appropriate privacy safeguards were in place.

As the report states, “Automated systems should be developed with consultation from diverse communities, stakeholders, and domain experts to identify concerns, risks, and potential impacts of the system. Systems should undergo pre-deployment testing, risk identification and mitigation, and ongoing monitoring that demonstrate they are safe and effective based on their intended use, mitigation of unsafe outcomes including those beyond the intended use, and adherence to domain-specific standards.”

Additionally, the administration suggests also considering building these systems to proactively identify and address potential areas of concern during design. Again, ensuring you are partnering with AI-model developers with a proven track record of following these best practices can help prevent any unwanted security and privacy issues from manifesting.

“Independent evaluation and reporting that confirms that the system is safe and effective, including reporting of steps taken to mitigate potential harms, should be performed and the results made public whenever possible,” adds the report.

In its announcement of the AI Bill of Rights, the White House made a special note to indicate individuals whose information may be exposed to AI-driven systems should demand the highest standards concerning how their data will be collected, transferred, accessed, used, and discarded. Only by taking proactive steps to build in the proper data safeguards at inception will you be able to provide the trust your clients want and deserve that government regulators will demand.

“Enhanced protections and restrictions for data and inferences related to sensitive domains, including health, work, education, criminal justice, and finance, and for data pertaining to youth should put you first,” notes the announcement. “In sensitive domains, your data and related inferences should only be used for necessary functions, and you should be protected by ethical review and use prohibitions.”

These types of consumer protections appear to be a top priority for the government, and those utilizing the technology will likely benefit from keeping them in mind as they look to gain the trust of their clients.

Trust is a critical part of the process

Two equally essential layers of trust are associated with implementing new AI-based tech solutions. The first layer of trust comes from organizations feeling confident that the tech stack they are working with meets industry standards. That, in turn, will ensure clients and systems users will also be confident in new work processes, which is the second layer of trust — the trust that comes with sharing this user experience with the rest of the world.

The development of AI systems and machine-learning technologies is ongoing, and there is little evidence that it will be slowing down any time soon. These are exciting times for businesses in every sector of the economy. Trust will be at the core of these new workflows’ success. At Thomson Reuters, we have designed products with this fact in mind, and for that reason, we have become a trusted partner for tax and accounting firms all over the world.

Conclusion

The tax and accounting industry is evolving in very exciting ways, and there are countless opportunities to use these burgeoning technologies to improve efficiency and bolster the bottom lines of tax and accounting firms. Embarking on this journey will present both challenges and opportunities, but with the right approach and providers in place to support you along the way, there is endless potential.

Therefore, we have redefined how companies work with the power of generative AI through our AI @ Thomson Reuters resource center. The arrows are all pointing toward a new wave of workflow technologies that will be developed and implemented inextricably alongside AI and its associated technologies. As firms continue to onboard, develop, and augment their generative capabilities, it will become increasingly important to have a firm grasp on how these technologies can boost your bottom line. Embarking on this journey will present challenges and opportunities, but with the right attitude, technology, and providers in place to support you along the way, there is great potential for explosive growth and a level of efficiency never before seen in the tax and accounting space.

Discover the future of professionals

As the landscape of industries undergoes a profound transformation fueled by the power of generative AI, it's imperative for professionals to grasp the potential and navigate the challenges that lie ahead. Our comprehensive report, "Future of professionals: Embracing generative AI," delves into the dynamic interplay between technology and expertise, offering insights into the opportunities that await those who embrace this paradigm shift.

Empower yourself with knowledge and insights that will shape your journey toward a future where AI and expertise collaborate to create unparalleled opportunities.

Got questions?

Ask Checkpoint.

With help from Checkpoint Edge, you’ll be able to take full advantage of emerging AI technologies to provide more efficient services for your clients