Tax Pros Using AI Can’t Keep the Human Element Out of the Loop

As artificial intelligence continues to pervade the regulatory landscape, tax firms and in-house tax organizations are beginning to worry about the use of AI in their own work—and in the results they generate for their businesses. Tax firms should prioritize elevating talent alongside AI implementation by developing their staff’s critical and strategic thinking skills.

Many tax professionals use AI through simple or complex prompts for chatbots. But AI likely will become increasingly involved in deep research and complicated analysis that once required human efforts. Agentic AI, which autonomously acts on them, is just one example of this.

Agentic AI can facilitate results with numerous details in a short time. But it’s still necessary to review their responses. Depending on the complexity of the tasks, it demands time to proceed with multiple layers of human checking. This is where it comes to the “human in the loop,” defined by Google “as a collaborative approach that integrates human input and expertise into the lifecycle of machine learning and artificial intelligence systems.” It’s expected that people are involved in the input and review of the results to ensure the accuracy of the output.

AI’s processed data can be based on billions of layers and knowledge sources, so humans may have a hard time digesting the massive amount of information it returns. This information includes tasks that cover qualitative and quantitative results, such as millions of transactions analysis, individualized compliance rules on multiple operations and jurisdictions, and deep research on international comparative tax legislation.

According to a recent survey from Ernst & Young, 53% of senior leaders say employees are “overwhelmed or exhausted by the constant influx of AI information and developments.” This trend is likely to affect the tax profession. Tax firms, in-house tax counsel, and their staff should consider the following questions for the future:

• Instead of giving guidance and orientation to humans, will they be giving prompts to agents/personas?
• Instead of managing interns and giving them opportunities to develop as professionals, will they be enhancing a tool’s machine learning abilities?
• How will interns and entry-level tax professionals get the chance to prove themselves?
• If the sector totally relies on AI for entry-level, how are we going to have tax experts in the future?

Companies should take this as an opportunity to instead elevate talent and invest in employee growth, whether through formal learning programs, hands-on experiences, or even using AI itself as a teacher of the firms’ own resources.

Tax firms and in-house leaders can rely on the available resources and the increasing support of AI tools to influence personnel on taking the time and opportunity to learn and apply them to their daily tasks—for example, organizing tax data results, providing comparisons on monthly tax compliance collection, and improving tax strategy communication for different audiences.

The Nobel Minds 2024 notes that AI will indeed take over many tasks and functions, but it won’t be able to replace human essence. Companies should train employees on soft skills that an AI or any automation could never replace.

Interpersonal skills, adaptability, strategic thinking, and critical decision making should be prioritized in careers for tax professionals to be able to manage AI well for it to work for them—not the other way around.

By improving employee training, tax firms will be able to add true value to the work AI delivers. Tax professionals should be able to formulate good prompts, review them, improve the conclusions, and efficiently communicate the results. This critical checking and reviewing process is what adds value to the conclusions, while AI delivers speed.

Leading tech companies, such as Uber, are giving opportunities for their employees to dive into AI possibilities and learn from formal and practical sources. Uber’s tax department created an internally focused cross-functional group to enable professionals from different backgrounds to participate.

The group aims to enable an environment of creativity on AI implementation by reuniting ideas, projects, and use cases worldwide. It also fosters diverse discussions through recurrent meetings on AI governance and internal development, as great minds don’t always think alike.

Tax professionals’ futures are heavily connected to companies’ abilities to create environments for investing in their employees’ careers and elevating talent, whereas AI implementation will have a greater role in the tax functions. Both matters must walk together to achieve their best professional and corporate potential. Let’s keep the momentum going.

This article does not necessarily reflect the opinion of Bloomberg Industry Group, Inc., the publisher of Bloomberg Law, Bloomberg Tax, and Bloomberg Government, or its owners.

Author Information

Flavia Sundfeld is a senior indirect tax specialist with the tax reporting strategy and planning group at Uber.

Write for Us: Author Guidelines