Board members and management are flooded with reports of how artificial intelligence is already – or will soon – impacting every area of ​​business. From supply chain and inventory management optimization to improved customer relationship management systems and dynamic pricing to personalized shopping and even optimized store layout, AI seems to be revolutionizing the way companies operate.

Amid the shockwaves, there’s one application that may not get the front-page attention as ChatGPT, but it’s giving businesses a lot of bang for their buck now, not 10 years from now: AI for tax. By helping tax departments keep track of various laws, meet compliance regulations, and prepare error-free tax returns, AI has already increased efficiency and accuracy—and it’s also giving companies a way to deal with a shrinking talent pool in tax and accounting.

Perhaps most important to boards of directors is technology’s ability to help them manage risk.

As Michael Charette, head of RSM Canada’s tax technology consulting practice, explained in an interview with Chief Executive Magazine, AI can identify potential risk areas in a company’s tax strategy by analyzing past audits and identifying patterns that have led to critical considerations in the past. By identifying these issues before they become problems, companies can proactively make adjustments and avoid costly fines and penalties.

For example, looking ahead to potential regulations involving real-time analytics of VAT and sales tax, companies need to prepare now while there is still time. “The lead time is not going to be as long as people think,” Charette said. “They’ll make an announcement and 12 months later you’ll have to comply. So it’s going to be really important to get ahead of this regulatory environment from a data perspective.”

The following conversation has been edited for length and clarity.

Given the abundance of issues on the board agenda, why should boards invest time and resources in tax automation?

From a talent perspective, it has become a necessity. There is a shortage of workers in the professional services sector in general. We are graduating fewer accounting graduates, and even fewer people are becoming CPAs (Certified Public Accountants) – and then even fewer People choose to do taxes after they become a CPA, so at some point all the money in the world isn’t going to get people to do them anymore because they just won’t exist anymore.

So you have two options: you can either outsource your tax department or you can automate it. A lot of that outsourcing is to other countries, and those countries are doing what robots would do if we took the time to build them. So the question for every professional services firm and the industry as a whole is: If we organize and clean up our data, if we have consistency within our organization, or if we make the effort to always know what is what, what could automation do for us?

And of course, the investment will pay off as technology costs come down. That’s the biggest challenge right now, getting a demonstrable ROI for automation in tax. If I have a tax department of five people and I spend, say, $500,000 on a particular process, it’s very hard to justify spending $200,000 on automation when I can do the whole thing for $500,000.

But then two of the five people leave, and now they can’t make it. And if they don’t make it, the penalties are millions of dollars. Then we have to remind them: It’s not just about full-time equivalents (FTE), it’s about risk management. The reason you need to automate is because you can’t scale your existing staff.

What best practices do you see among companies driving automation?

First, there should be a company-wide strategy for data, and taxes should be part of that strategy. Leaders across the organization, including at employee level, should have a certain level of digital and data literacy to understand what is possible.

And I think something has to change when we apply Six Sigma, Toyota’s culture of continuous improvement, to the world of finance. We always have to ask ourselves: is there a better way? If only for our own sanity, because in the digital age we have a lot more data at our disposal. Professional accountants absolutely have to be like budget data scientists to some extent, so those skills have to be there. But leadership also has to give people the opportunity to experiment – in a controlled setting so they don’t take risks.

So best practices include having a data strategy, making sure tax is part of it, improving data literacy, and creating a culture where people can try to improve efficiency internally – because it costs too much to hire someone like me to fix every little thing.

What can board members ask management to do to ensure this problem is addressed?

First of all, what do we do with the data? What is our data strategy at the enterprise level? Risk management data should be part of it.

Data is a critical corporate asset, so there should be a plan for how we grow that asset. How do we manage that asset? How do we protect that asset from risk? Leadership doesn’t need to know the details, but they should have a plan, and someone should be accountable for that plan.

Data isn’t worth much if it’s not used. That’s kind of the next challenge: How are you going to use that data and for what? Companies already know that. They’re already doing it in the front of the house. Customer analytics is nothing new. But back of the house business decisions – even investment decisions – aren’t always based on the best data. There’s a lot of assumptions and a lot of guesswork. Technology is there to make it better.

Companies now have employees working from all over the world. With greater access to this high-quality data, can they, for example, better model the tax implications of new hires?

It could – and there are other ways to save costs. Global mobility is a constant issue. Whether I’m an engineer or another professional or managerial type, if I have to work somewhere for a month, there are tax implications. If I’m a professional basketball player and I play for New York but have a game in California, California wants that pound of flesh for that game, right? That’s just the reality. Employee self-reporting is really unreliable and it’s unfair to employees to expect them to remember all the data.

For example, for one client, we did an analysis of employee movement for state and local taxes in one area and then global taxes for employees moving across Europe, Africa and Asia. We realized we had data that indicated where everyone was. So we started talking to their logistics team about planning projects. What they didn’t realize was that they had an employee moving from China to Europe, Europe to China and China to Europe – but they already had the same skills in Europe. So the cost of moving that employee back and forth was huge – while they had a resource in the bank in Europe that they didn’t know about.

The insights were in the data, but nobody was going to look through a huge spreadsheet and filter them out. They had to see them visually because color, shape and size give context. When they saw them on a map, when they saw the little green dot that represented this employee and they hovered over it and saw what skills they had, they thought, “Oh dear, why are we wasting so much money sending this person back and forth?”

Honestly, we just wanted to highlight where everyone stands and what their tax obligations are. But if taxes are part of your data set, you can base other business decisions around it.

We all hear the scary warnings about AI and the potential risks of technology replacing humans. Why shouldn’t boards be concerned about the use of AI in tax?

So there’s generative artificial intelligence, like what ChatGPT is doing at the forefront. Then there’s very purpose-built, specific artificial intelligence and machine learning, which is basically a statistical product. It says, “There’s a high probability that this is the answer based on all the inputs you’ve given me.” Applying that to taxes is pretty significant.

We often ask tax people for their opinion, and tax people are good at being aware of the problem, they know the court cases, they know the laws, they know how to find information, they know how to write a memo. But AI has been trained in everything. AI can be trained 10 times, and once it’s trained, it can train another AI 10 times, then another, and so on – it can learn exponentially.

That doesn’t mean we’re excluding humans. It just means that the role of humans is no longer to search for information, coordinate it, summarize it, and compile opinions. It means that the machine does a first pass and says, “Here’s everything I found on this topic – do you like it? Yes? OK, here’s a summary of all these thoughts, a summary of everything that’s happening there. Do you like it?” And then you, as a human, have to explain what it is. But we’re not talking about a generative intelligence that can push you out of that role.

On the machine learning side, there’s a lot of value in just saying, “Hey, there’s a statistical probability that this is happening because of all these other things.” Companies exist within jurisdictions; they work with products and they work within communities. So there are all these little factors that affect tax consequences. If we can build machine learning models that take all of these factors into account, we can get better predictive results based on them. But there’s no point where the AI ​​says, “Hey, wait a minute — that’s illegal. I’m calling the IRS.”

As finance is transformed – and every transformation these days is about data and data strategies – we are increasingly seeing tax leaders sit at the table from the beginning and be able to provide input. “How do we configure the enterprise resource planning (ERP) system? How do we configure our data warehouses? How do we make sure that all the systems that our ERP and data warehouses are connected to are actually thinking about tax, so that when the tax people do their work, the data is actually in a single source of truth?”

That’s still the challenge of getting every organization there. But once that rich data source is aligned with tax requirements, tax leaders have to get more creative and say, “OK, we can automate all of this, but can we do more with it?” That’s the next evolution.

Read the full report>>

Print article