The ‘robot uprising’ and you: A CPA leader’s view on harnessing AI

This week’s episode is a mix of CPA-centric news and a Q&A with a leader in the profession. First is a summary of the House Ways and Means Committee’s tax bill, the AICPA’s reaction to some of its provisions, and an update on paths to CPA licensure.

Danielle Supkis Cheek, CPA, is the guest on the main segment, discussing how accountants can avoid head-spinning dizziness when it comes to artificial intelligence (AI) — and why the profession is well suited for technology adoption.

Supkis Cheek, senior vice president–AI, Analytics and Assurance at Caseware, also discusses how organizations can use AI to scale operations. Finally, citing a set of criteria from Digital CPA speaker — and JofA podcast guest — Alex Dorr, Supkis Cheek rates the amount of drama she brings to the workplace.

What you’ll learn from this episode:

• A summary of top JofA news items from earlier this week.
• Why conference speakers on AI topics have a hard time providing timely slides for sessions.
• The importance of the evergreen nature of accounting standards.
• Supkis Cheek’s explanation of professional skepticism’s role in harnessing AI.
• Detailing the public expectation gap as it relates to audits and technology.
• Why Supkis Cheek considers the AI era a curation age instead of a technology age.

Play the episode below or read the edited transcript:

— To comment on this episode or to suggest an idea for another episode, contact Neil Amato at Neil.Amato@aicpa-cima.com.

Transcript

Neil Amato: Welcome back to the Journal of Accountancy podcast. We will get to the interview with Danielle Supkis Cheek in just a bit, but first we have a few news highlights.

On Wednesday morning, the House Ways and Means Committee approved its proposed tax bill. The bill would extend or make permanent several key aspects of the 2017 Tax Cuts and Jobs Act. It would eliminate taxes on some tips and overtime pay and create a “senior bonus” deduction, among many other provisions.

The Journal of Accountancy’s coverage details the bill, including those parts the AICPA endorses and those with which it has concerns. We will link to that coverage in the show notes for this episode.

Also on Wednesday, the AICPA and NASBA, the National Association of State Boards of Accountancy, jointly announced the approval of changes to the Uniform Accountancy Act. Those changes add a new path to CPA licensure. Bryan Strickland has that coverage, also linked in the show notes.

Visit journalofaccountancy.com for the latest on these and other topics. Now, here’s the interview with Danielle Supkis Cheek. She had just presented at Digital CPA in Denver when we spoke.

Back on the Journal of Accountancy podcast, this is Neil Amato with the JofA. A repeat guest is on the show for this episode. Her name is Danielle Supkis Cheek. Danielle, you’ve just come off stage. We’ll get to that session in a bit, but first, thanks for being back on the podcast.

Danielle Supkis Cheek: Thank you for having me again.

Amato: We’re glad to have you again. You and Jim Bourke presented a session with this title, “Audit Technology for Tomorrow: Current and Emerging Tech in the Audit Space.” First, I want to ask, for such a fast-moving topic, how challenging is it just to have an updated slide deck?

Supkis Cheek: It’s almost impossible. But this is all of my slide decks. Especially in the world of AI, by the time you finish your deck and especially a lot of conferences want them weeks if not months in advance, the tech has totally changed. I like to say it in doggie years concept that a quarter in AI is a real-world one year, and so anytime you ask me to submit slides six weeks, two months in advance, it’s the better part of a year in advance. The tech is changing that fast.

Now, there’s some core pieces and foundational concepts. When you start to think about principles and how to think about very metacognition, that stuff is actually pretty standard, pretty basic. If you think about the standards themselves of what we have to abide by as a profession, that stuff is very evergreen and lasting.

The pieces that I try to always submit are very theoretical about how to think about problems, and I try to focus less on the specific tech because I know the specific tech will continue to change and innovate over time, and I’m never going to be able to keep up. Nobody can actually fully keep up. If we always keep it back to the principles of what we are expected to do, then not only does the slide deck last longer – I can make it those two months – but it also means that the concepts that I’m teaching people of how they can think about problems will serve them better in the future.

Amato: I didn’t say it before, but Danielle’s title at Caseware is VP of analytics and AI. Not that that matters other than making the proper introduction. But I want to ask you about this tenet of your session. Methodology is not the problem in technology. What do you mean by that?

Supkis Cheek: It used to be that I used to hear on a really regular basis, “The standards don’t allow me to use technology, they’re not either prescribed, they’re not written into, they specifically even prohibit” – is a lot of the statements I used to get, and this was a couple years ago, so in ancient times. Sorry, I can’t help myself.

Amato: A quarter is a year ago, so a couple of years, that’s like?

Supkis Cheek: Eight years ago.

Amato: Eighteen quarters.

Supkis Cheek: But I think everybody’s perception, or most people’s perception, is that the standards themselves do not prescribe a certain kind of technology. They don’t even prescribe many times the how to execute. We as humans have biases, and we are very biased in our traditional and historic approaches because they’re tried-and-true. They have worked for us. Even when you start to think about new ways to solve the problem, there’s still an element of do the standards allow it because of our interpretation of how to execute. I really like to have people that have never read standards before go back to the original pronouncement of something and read it. They will come up with a completely different interpretation than the way it is today.

If you start to think about fresh reads and go back and see what is almost the intent of what we’re trying to cover from a particular standard, it completely reshapes your understanding of what is permissible, how to execute even, and that’s where the technology can start to layer in and ends up being the how do you execute? The technology is just a different tool on how to execute than the ways of the past. It’s not anything that is overly new and fundamental in the sense of we’ve had technological innovation of the past.

Economics talks about economic churn, and the highly cited example is always going from horse-drawn carriages to the Model T. There’s been these constant shifts in technology. It’s just happening faster now, and it’s happening in such a fast way we’re not even making it through certain audit cycles. That’s what creates the uncertainty and the complexity, but it’s not the rules that prohibit you in the accounting profession. Actually, the accounting profession is very well suited for technology because we have such principally based standards, and I know I’m going on, but this is one of my favorite topics, as you know, I can’t get off, let’s keep with high horse about this.

But if you think about the concepts of not subordinating our judgment to others and that includes the robot uprising, our profession has that professional skepticism in it and those ethical requirements that make us be independent thinkers, that make us think through how we’re executing, think through how do we supervise appropriately or oversight properly technology. That’s actually going to serve us really well for the fast-paced innovation of the future and make the correct amount of reserved and guarded on “newfangled tech,” but give us that healthy skepticism of making sure it’s safe and making sure we can rely on it appropriately and not getting overly caught up in the flash bells and whistles shiny object, but actually the practical application of how do I comfortably from a standards and safety perspective, execute this particular task, and how can I do it in a new, innovative way that is still safe?

Amato: Safe but fast. I don’t even know if this is the right concept, but one of the things that was mentioned in your session was the public expectation gap as it relates to technology. What do you mean by that?

Supkis Cheek: To me, that’s the ultimate why. Why do we care about technology? I think the post-Wirecard world of audit tech and audit approaches is going to be as change-driven as the post-Enron world of audit theory. Wirecard changed the fundamental understanding or showed the fundamental disconnect, if you will, of what is the public’s expectation of what an audit is and the actuality of what is defined as reasonable assurance. That public expectation that we are NCIS or whatever fancy crime TV show that has provided absolute assurance.

That’s what the world believes audit to be: absolute assurance. We all know, your listeners know, reasonable assurance, that’s all we’re responsible for; we can’t really ever get to absolute assurance. That is an unwinnable target. Things will always be moving on us, so that what is in the middle between those two is considered the expectation gap. There’s many ways to different close it. But if you start to think about what is the regulators, the standard setters, and the holistic public interest that we all serve is having an expectation of more from us, in theory. They want more. They want that reduction in the expectation gap.

We still have our standards. We are still defined as reasonable assurance, but what is reasonable will always be moving on us because of what is normal for today. It’s based on what is reasonable, what is normal for today. As you start to think about, the public expects more out of the profession, and yet we are in a position where we actually have a lot of us are looking at less resources coming to us. I’m sure you’ve had people talk about pipeline before, so where does this all meet?

This is where the rubber meets the road, and so technology is going to be one of the ways that we can help scale in a more meaningful way. AI is going to clearly be an important part of that. But if we think about why do people care so much about that technology, it’s because we have to be able to keep up with the times, and the times right now are everyone expects more with us and for many of us, we’re working with less resources. How do we bridge that? Technology is going to be part of the how. It is not the silver bullet. It’s not going to be the only way, but it’s going to be part of it.

Amato: Another phrase of yours, curation age. That’s the age that’s either here or coming. How is AI playing a role in that?

Supkis Cheek: Yeah. I don’t like to think of this as the AI technology age. I think, again, that’s just the how do we execute? If the past was the information age, and the information age has given us so much rich, valuable content. It is almost impossible to know things in real time contemporaneously in a meaningful way when you’re looking for just something being off. It’s going to be our ability to curate and find the information that is relevant. The early keynote speaker this morning, she was great, talked a lot about synthesizing information and summary information from AI. Absolutely great use case. She was all about it, rightly so.

It’s the concept of the next layer of information is curation and being able to take what is really complicated, simplify it down, find the needle in the haystack, find the pieces of information in a faster more efficient way. Because guess what? With hindsight and the volume of data that is there, everybody’s going to be able to be in a position where a couple of years down the road, you’re able to say, I told you so. You should have known better because the data was there.

Well, guess what? When you’re in real time, you don’t necessarily know those things. It’s how can we curate that information as quickly as possible for whatever is our relevant task. I think that’s going to be the approach. Yes, AI is a big technology, and how do we curate? But I really think the core underlying capability is it’s a curation concept rather than just it’s an AI age, because AI is just a how-to-execute approach, and there’s many different types of AI obviously on how you could execute.

Amato: Are you going to be VP of curation?

Supkis Cheek: I don’t know.

Amato: You said AI is just an age. Well, it’s in your title.

Supkis Cheek: It is. I guess I do work at a technology company, so I do have to focus on what kinds of tech I provide and I’m responsible for. But yeah, I love it. I’ll bring that one back to the bosses that be.

Amato: Is there anything that you’ve come away with from another session? You mentioned the early keynote on this Tuesday that we’re recording. Anything else that you’ve come away from Digital CPA as something you’ll take back?

Supkis Cheek: Alex Dorr yesterday, amazing session. Really fundamentally changed my understanding of how to think about the human talent and his equation on the human experience and how balance works. I love mathematics. I’m a math geek. I love things balancing out, I love a good double entry. He has a zero point in there, because that’s the perfect point of balance. I don’t think I’ve ever thought of myself as equal, not equal in the sense of I think of myself as a high performer. But yet, I have a lot of needs. I work fast. I put a lot of stress on teams. Like, that creates an opposite effect on it.

Amato: Do you work as fast as you talk, by the way?

Supkis Cheek: Yes, I do, but maybe faster. I need another set of hands to type faster. If I could, I would. Oh, I think so fast, it’s not even funny. My brain is a scary place, but that’s why I love these sessions that Digital CPA always hosts because I wasn’t coming here to think about that issue, but it changed. For me, the human-in-the-loop process to supervise and oversight AI is so important. It’s what actually makes it so that a lot of those that are still using legacy audit processes are actually better, well suited for the current change management, because they didn’t go down the path of hyperautomation.

Their processes already have a heavy human review element. It’s great. It’s perfect. It’s well suited for the change management. To think about humans in a bit more balanced to the way and in a place that I don’t think he meant to liken it to double-entry accounting, he didn’t actually make the analogy, but his equation, I love a good equation. But then it keeps that balance. It was great. It was just absolutely wonderful. Going to love to take that one back to the office and do a little bit more learning about his approaches. That was a great one. Sorry, I’m just gushing at this point. I should stop gushing.

Amato: No, that’s quite all right. We liked Alex enough that we had him on the podcast for the first time.

Supkis Cheek: We’ll put a link to that one in this one.

Amato: Yes. We’ll see which one airs first. I’m not sure yet, but it’s a great point. What I will ask, though, since you were in his session, are you a high-drama employee?

Supkis Cheek: I put myself at a two on the drama. I did. We were arguing a little bit of, am I two? Am I a three? I think I’m a two. It’s actually a little bit hard to grade yourself on these things, and I don’t have a lot of my team members at my table to grade me.

Amato: I don’t actually know what’s good on that scale.

Supkis Cheek: One is the best. That you’re low-drama. But in my role, I think I need a little bit of drama and excitement to really drive. Three is still considered a normal status quo. I think I’m, and this may be my own bias. I talk about biases all the time. I’m biased of myself. It’s hard to accurately grade myself. I think I’m less drama than everybody else, but probably my team would suggest otherwise. So depending on who’s listening.

Amato: That’s right. We’ll see. Danielle, thank you very much. We appreciate having you on again.

Supkis Cheek: Always love being here.