Earmark CPE

Earn CPE Anytime, Anywhere

AI

The Accounting Profession Has AI Completely Backwards

· ·

When Accounting Today surveyed industry thought leaders about AI’s impact on the profession, every expert agreed that AI would automate the boring stuff like bank reconciliations, data entry, and transaction matching while humans would rise to strategic advisory work. Not one thought their own job was at risk.

On a recent episode of The Accounting Podcast, hosts Blake Oliver and David Leary did something clever. They fed the same questions to ChatGPT, asking it to respond as an accounting thought leader. The AI’s answers were just as good as the human experts’.

“None of the accounting thought leaders think their job could be replaced,” David said, “which is crazy because essentially AI can at least do the thought leader job.”

Blake and David argue that the profession has AI’s impact exactly backwards. While everyone confidently predicts automation will eliminate mundane bookkeeping tasks, the technology actually excels at synthesis, narrative-building, and strategic analysis—the very work that defines “thought leadership.”

What AI Actually Does Well

The standard story about AI in accounting is machines will handle the boring, repetitive tasks while humans ascend to strategic advisory work. It’s comforting and logical. But according to Blake and David, it’s completely wrong.

“AI can take financial statement information and turn it into a narrative better than I can, better than almost anyone can at this point,” Blake states. “That’s what we should be using it for.”

Consider Mike Salvatore, a Chicago business owner with two cafes, two bars, and a bike shop. He used to analyze his cost of goods once or twice a year, spending hours crunching numbers. Now he does it every three weeks by feeding data from QuickBooks and his point-of-sale system into Google’s NotebookLM, which creates a podcast-style summary of his business performance. He sends these AI-generated recordings to his managers.

“It’s essentially my CFO,” Salvatore told The Wall Street Journal.

This isn’t AI doing mundane bookkeeping; it’s performing executive-level analysis and communication.

Blake’s own experience drives the point home. He built an AI system that turns news articles into detailed research notes and social media posts. That work used to eat up hours each week. He also trained an AI ghost writer on hundreds of his past writings. Now he can dictate a voice memo and get back a polished article in his own style.

“Basically, it has made it so, as ‘thought leader,’ I don’t do any of that anymore,” he admits. “It’s like I have a team that does that for me. I started working out and I’m just enjoying life.”

Meanwhile, the supposedly “easy” transactional work is stubbornly resistant to automation. David, who spent years taking QuickBooks support calls before co-founding the podcast, gets fired up about this misconception.

“Matching bank feeds is not bookkeeping. That’s just matching,” he argues. “Accounting is sending an invoice to somebody so they’ll pay me.”

He describes his recent struggle trying to upload an invoice to a client portal. It’s a “mundane” task that should be simple but isn’t. The process requires navigating confusing interfaces, making contextual decisions, and handling exceptions that don’t fit predetermined patterns. AI can’t do this reliably because it lacks the real-world context that humans take for granted.

The disconnect is striking. Thought leaders keep repeating the same message they’ve preached about cloud accounting for a decade: technology will free you up for advisory work. But as David points out, “I don’t think AI is freeing up your time to do that work yourself.” Instead, AI is doing the advisory work directly.

Are You Willing to See the Opportunity?

Where things get interesting is the same AI capabilities that threaten thought leaders create a massive opportunity for regular practitioners if they’re willing to see it.

Mike Salvatore, the Chicago business owner interviewed by the Wall Street Journal, wasn’t working with an accountant before. His AI “CFO” didn’t displace a human. He simply started getting insights he’d never received.

“Very few accountants serving Main Street businesses will actually do that kind of work for a price these business owners want to pay,” Blake explains. “So they do it themselves, but they don’t do it often and they don’t do it well.”

AI is filling a vacuum, not replacing existing services. And that vacuum is huge.

If a business owner can get advisory insights that are even 50-80% accurate from AI, that’s better than the nothing they’re getting now. The question for accounting firms is whether to let clients figure this out themselves or to offer AI-powered advisory services with professional oversight.

“Firms can feed data from clients’ QuickBooks files and their point of sale systems into these tools to generate AI analysis,” Blake suggests. “You can charge for it, because you’re adding the oversight—checking the numbers, making sure it actually makes sense.”

David connects this to a decade-old challenge. He remembers when LivePlan tried to train bookkeepers to offer business planning services. “They really struggled with it because they’re good at bookkeeping. But it’s hard to teach somebody to tell a story and create the narrative around the numbers.”

Now, “all those bookkeepers can basically offer that with AI out of the box and charge for that additional service.”

When ChatGPT (playing the role of thought leader) was asked what would make it worry about being replaced, it gave a revealing answer: clients accepting “AI-generated advice as good enough, even in ambiguous scenarios.”

Blake’s interpretation is blunt. “That’s what AI will fill—the gap in the market where accountants aren’t providing the service. There’s a big gap and there aren’t enough of us.”

Why Billable Hours Kill Innovation

One survey question asked about the “AI premium.” How much more should an AI-savvy accountant earn compared to an identical colleague who doesn’t use AI? The thought leaders said these employees should obviously be paid more.

Blake laughed at this. “How can you pay them more if you’re looking at them in terms of billable hours? AI is going to actually reduce their billable hours, not add more.”

If an employee uses AI to finish work in half the time, they bill half the hours. Under the traditional model, they look less productive, not more. Under the traditional model, “you should pay the AI employees less because they’re working less,” Blake points out.

This creates a ridiculous situation where your most innovative, efficient employees appear to be your worst performers.

Ryan Lazanis, who built and sold an accounting firm and now coaches other firm owners, has a different approach. He focuses on just two numbers: bottom-line profit and monthly recurring revenue. Not billable hours, utilization rates, or time per client.

“He is not breaking it down by client. He’s not looking at individual job profitability,” Blake explains. The only thing that matters is whether the firm made money over the year.

This makes sense because staff costs are fixed. “The amount of hours they spend has no impact on your profitability,” Blake notes. You only need to worry if one client is so demanding they prevent you from taking on others.

“You don’t have to track hours for months to figure out which clients are eating up your profits,” David adds. “You just go to your team and say, ‘Who’s the biggest pain in the ass client?’ And they’re going to tell you.”

There’s also a technical angle to consider. Blake cites research showing AI is nearly 100% accurate on tasks that take humans 4-5 minutes. That accuracy drops for longer tasks, but the threshold is “doubling every seven months.” By the end of 2026, AI might handle 10- to 20-minute tasks reliably.

But this only matters if firms can capture the productivity gains. Under billable hours, faster work just means more hours to fill. Under outcome-based metrics, faster work means more capacity for growth.

Is the AI Accounting Influencer Coming?

As the episode wraps up, Blake and David float an idea that captures the absurdity of the current moment. They’re considering creating an AI accounting influencer—a completely artificial thought leader to see if it can build a following comparable to real industry voices.

“Let’s make an AI accounting influencer and see if we can build its following to eclipse that of those real influencers,” Blake suggests. They could have it write newsletters, create content, maybe even land sponsorship deals.

It’s partly a joke, but it makes a serious point. If an AI can answer thought leadership survey questions as well as humans, write articles, and provide strategic insights, what exactly makes human thought leaders irreplaceable?

The answer might be less comfortable than the profession wants to admit.

Looking Ahead

The Accounting Today survey offered some important insights, though probably not what it intended. The people most confident about AI’s limited impact are those whose work AI does best. When ChatGPT generated answers indistinguishable from human experts, it demonstrated the very vulnerability those experts deny.

The real story is that AI excels at synthesis and narrative, which are the heart of advisory work, but struggles with the contextual, exception-filled world of everyday bookkeeping.

Firm owners should rethink their services to capture the advisory opportunity AI makes possible, and abandon billable hours before they strangle your ability to innovate.

For individual practitioners doing transactional work, the news is actually good. Your skills remain valuable precisely because your work requires the messy, contextual judgment that AI lacks.

And for thought leaders? As David observed with obvious frustration, the elitist attitude that “I’m better than you” has been in accounting for 30 years. “The reality is completely opposite. People are completely missing what’s really going to be replaced by AI.”

The race isn’t between humans and machines. It’s between practitioners who recognize AI’s true capabilities and those who cling to comfortable narratives while missing the transformation happening around them.

To hear more about Blake’s AI-powered lifestyle, David’s thoughts on what bookkeeping really is, and their plan to create an AI influencer that might outperform the human ones, listen to episode 469 of The Accounting Podcast.

The Auditors Got Red Flags About $95 Million in Missing Funds and Signed Off Anyway

· ·

In the last episode of 2025 of The Accounting Podcast, hosts Blake Oliver and David Leary kicked off the conversation with an unexpected problem: America is running out of pennies. David’s friend owns sandwich shops in Tucson and literally can’t get pennies from the bank anymore. Businesses are being forced to round to the nearest nickel, and point-of-sale systems are scrambling to adapt.

“Square admits one fifth of all the transactions on Square are still paid in cash,” David noted, highlighting how this seemingly small issue affects millions of daily transactions. The government claims there are 300 billion pennies in circulation, but as David pointed out, “Obviously this isn’t true because businesses all over America do not have pennies to use in transactions.”

But the penny shortage was just the warm-up. The hosts quickly moved to a much bigger story about missing money: $95 million vanished at Evolve Bank, yet the auditors still signed off on clean financial statements.

$95 Million Went Missing While Auditors Said Everything Was Fine

Blake followed the Evolve Bank story for years, and recent Freedom of Information Act requests uncovered stunning details about what the auditors knew and ignored.

Evolve Bank is a chartered bank that worked with Synapse, a “banking as a service” company that wasn’t a bank itself but managed the technology connecting consumer apps like Yotta and Juno to actual banks. When you used these apps, you’d see your balance, but you had no idea which bank actually held your money. Synapse managed all those details.

“Everything worked great until April 2024, when Synapse filed for chapter 11 bankruptcy and shut down operations,” Blake explained. Suddenly, the banks and the apps couldn’t figure out where customer money actually was. Evolve froze withdrawals from thousands of accounts, leaving people unable to access their own money for months.

When banks examined Synapse’s records, they found massive problems. Between $65 and $95 million in customer funds couldn’t be traced to any bank. “Your Juno account might say $10,000, and the bank that’s supposed to have the $10,000 says, ‘We don’t have it,’” Blake explained.

The most damaging revelation came from the 2023 audit. When Crowe, Evolve’s auditor at the time, asked Synapse to confirm cash balances, the response should have triggered immediate action. Evolve listed 113 accounts, but Synapse was missing 29 accounts from daily data feeds. Synapse’s general counsel asked to discuss the discrepancies with Evolve’s leadership.

Evolve never responded to that request, yet Crowe still issued a clean audit opinion.

“Ninety five million is a lot of money,” Blake observed. “It would be 6% to 7% of Evolve’s total assets, likely over 100% of their annual net income, a double digit percentage of equity capital in some years. And typically, materiality would be 1 or 2% of assets.”

Are SOC 2 Reports Worthless?

The Evolve disaster led the hosts to question other compliance frameworks, particularly SOC certifications that companies display as badges of trustworthiness.

“My guess is Synapse had their SOC 2, because it’s not that hard to get a SOC 2,” Blake said. “According to my understanding, it’s really just a lot of documentation of the controls. But there’s not necessarily any confirmation that those controls are being followed.”

“They paid the money and got the badge for their website,” David observed. 

The hosts also discussed how a New Jersey accounting firm, Sax, took 18 months to inform nearly 250,000 people about a data breach. The firm claimed it followed standard procedures and saw no evidence the stolen data was misused, but for 18 months, affected individuals had no idea their personal information might be compromised.

“People could be using your stolen identity fraudulently 18 months before the accounting firm lets you know,” David said.

The problem is that while firms must have Written Information Security Plans (WISPs), they’re not necessarily legally required to execute them properly. “We focus on the wrong thing,” David argued. “We focus on having a WISP, not actually executing the WISP.”

Partners Don’t Know What Partners Make

In a lighter but equally revealing segment, Blake shared his favorite LinkedIn post of the year from Chase Birky, CEO and Co-founder of Dark Horse CPAs. Chase shared that almost a third of partners don’t know how much partners make at their own firms.

“How do CPAs not know how much they make? Isn’t that sort of what we do?” Chase wrote. The problem stems from a lack of transparency at many firms where partner compensation is calculated in a “black box” and communicated well after the fact. This secrecy is at least part of the reason talent leaves public accounting.

“I left because I got offered a job in tech that paid a lot more,” Blake said, sharing his own experience. “I didn’t know how much a partner made, and nobody could tell me what the path looked like.”

Should Companies Report Twice a Year Instead of Four Times?

The hosts also debated President Trump’s suggestion that U.S. companies should move from quarterly to semi-annual reporting, like much of the rest of the world.

Research from the UK showed that changing reporting frequency had “virtually no impact on companies’ internal investment decisions.” Studies also found that quarterly reporting creates “noisier data” that benefits sophisticated investors while hurting everyday investors.

“We’re always talking about how we have too much work to do in accounting and we’re pressed for time,” Blake said. “What better way to give ourselves more time than to make it two times a year instead of four?”

David wondered if less frequent reporting might reduce the pressure to play accounting games. “Monthly reporting probably puts pressure on people to sidestep the rules because it’s so fast and you have to perform.”

Looking Ahead to 2026

The hosts wrapped up with predictions for the coming year. David was skeptical about AI transforming bookkeeping. “I don’t see me doing bookkeeping at the end of 2026 any differently than I did in 2025, 2024, 2023, or 2022.”

Blake disagreed, pointing to new AI browsers that can actually navigate accounting software and complete tasks. “The time is doubling every seven months,” he explained. “Within the next year, we’re going to see AI able to complete tasks that take 15 to 20 minutes with 100% accuracy.”

David also predicted that OpenAI would strike a multi-million dollar deal with the AICPA, that at least two AI companies would fail, and, in his easiest prediction, “Intuit will tick off accountants in 2026.”

The episode covered far more ground than can be captured here, from the technical details of audit failures to the future of AI in accounting. For the complete discussion and all the insights Blake and David shared in their final episode of 2025, listen to the full episode of The Accounting Podcast.

The End of Data Entry and What It Means for Your Tax Practice

· ·

Elizabeth Beastrom left public accounting 30 years ago because she was sick of rekeying data into tax returns. Now, as President of Tax and Accounting Professionals at Thomson Reuters, she works to make sure no accountant has to do that mind-numbing work ever again.

“I was a lazy CPA,” she admits with a laugh during this episode of the Earmark Podcast. “I didn’t want to spend my time doing work that I didn’t think was necessary.”

In this conversation with host Blake Oliver, Elizabeth and Kirat Sekhon, Thomson Reuters’ Head of Technology, map out their vision for automating the entire tax workflow, from gathering documents to delivering returns. They want listeners to know that AI-enabled firms are going to outcompete everyone else, and the shift from compliance to advisory isn’t optional anymore.

Why Tax Firms Can’t Keep Doing Things the Old Way

The numbers tell the story. Fewer people are taking the CPA exam while more accountants retire every year. Meanwhile, tax complexity keeps growing, which means more demand for services with fewer people to do the work. Throw in private equity firms buying up practices and pushing for efficiency, and you’ve got a perfect storm.

But it’s not just about headcount. The new generation of accountants expects modern tools that actually work together—not the clunky desktop software their predecessors put up with.

“They expect to use intuitive and connected tools,” Kirat explains, “so they have a better experience while they deliver value to their customers.”

So why has tax software stayed stuck in the desktop era while cloud accounting tools have taken off? Kirat points to two reasons. First, tax calculations are hard to get right, and once you build something that works, nobody wants to break it. Second, accountants themselves haven’t pushed for change. When you’re working 80 to 100 hours during busy season, the last thing you want is to learn new software.

“The term SALY—same as last year—still comes through,” Elizabeth notes. “You found a way to do it and you like to replicate that. Change is hard, especially when you have to introduce that to the firm when you’re working 80 to 100 hours a week.”

But resistance to change is becoming dangerous. Elizabeth’s own exit from the profession 30 years ago shows what happens when the work becomes too tedious. Back then, she discovered she loved the advisory side, including talking to clients, understanding their businesses, and making recommendations that actually helped them improve. But she was stuck doing data entry.

“I would spend time talking to my customers,” she recalls. “Some of my best inputs came from the people in accounts payable or accounts receivable. I would get a detailed understanding of their process.” But then she’d have to go back to rekeying tax data, and the contrast was too much.

Building the “Bookends” Around Tax Prep

Thomson Reuters isn’t trying to fix one piece of the tax workflow; they’re automating the whole thing. Their strategy focuses on creating what Elizabeth calls “strong bookends” around their core tax engines (GoSystem Tax, CST, and UltraTax).

The front bookend came through their acquisition of SurePrep three years ago. Practitioners dump all their client documents into the system, and SurePrep automatically classifies them, pulls out the relevant numbers, creates a binder for review, and fills in the tax software. No more manual data entry.

“That’s a huge time savings when you don’t have to spend time doing all of that manual data entry,” Kirat says, “and they can actually focus on the return.”

The back bookend arrived with SafeSend, acquired earlier this year. It handles return delivery, e-signatures, and payment collection, eliminating what Elizabeth remembers as the nightmare of printing, mailing, and faxing documents back and forth 30 years ago.

What’s different about Thomson Reuters’ approach is they’re keeping these tools open to work with competitors’ software too, not just their own tax products.

“It is an open, curated ecosystem,” Elizabeth emphasizes. “If customers find value in part of their workflow, we want to make sure we connect to it.”

Beyond just automating existing steps, they’re trying to eliminate unnecessary work entirely. Take the client questionnaire—that paper organizer Blake’s mom still fills out by hand every year. Thomson Reuters wants to “kill the questionnaire” by using AI to pre-populate information from prior returns and only ask for what’s actually new or missing.

The next frontier is what Kirat calls “agentic AI,” systems that don’t just handle one task but orchestrate entire workflows. These AI agents can use multiple Thomson Reuters products in sequence, making decisions along the way to get a return from start to finish with minimal human intervention.

But everything the AI does needs to be auditable. Kirat stresses that any AI handling tax work must show exactly what decisions it made and why.

“Our customers expect the work product of an accountant to be 100% accurate,” she explains. “Without providing that audit log with the decisions and choices and confidence levels, we’re missing the mark.”

Blake agrees enthusiastically, sharing his frustration with current AI tools that don’t show their reasoning. “I want to know why it matched this transaction,” he says. “There’s an AI conversation for each one of these transactions. Why not give that to us?”

The Shift to Advisory Can’t Wait

If machines can prepare returns faster and more accurately than humans, what exactly are clients paying for? Two-thirds of Thomson Reuters’ customers say they want to shift to advisory services, but most don’t know how to actually do it.

Enter Ready to Advise, launched in June 2024. The tool takes everything from a completed return and analyzes it against potential tax strategies based on that client’s specific situation and goals.

“It will quantify the savings,” Elizabeth explains. “It will ask for more information to get to a range. It will allow you to have that discussion where you can say, ‘Hey Blake, I noticed from your 1120-S filing some potential strategies you should take.'”

Then it walks you through implementing those strategies and produces client-ready documentation. For firms struggling to move beyond compliance, this is huge.

But technology alone won’t fix the business model problem. Clients have been trained to expect strategic advice for free. “I might call my accountant and say, ‘Hey, tell me what this big beautiful bill does for me this year?,’ which is code for don’t charge me for this,” Elizabeth says, capturing the conundrum perfectly.

That’s where Practice Forward comes in. It’s Thomson Reuters’ tool for helping firms understand their worth and develop advisory pricing models. The goal is shifting from hourly billing for returns to year-round advisory subscriptions.

Ready to Advise also solves a talent problem. Traditionally, you needed years of experience before you could offer meaningful tax advice. But with AI assistance grounded in Checkpoint’s content (maintained by over 4,500 subject matter experts), newer staff can contribute to advisory work much sooner.

“That junior associate’s experience, paired with all the knowledge that there is available in generative AI today, is incredibly powerful,” Kirat notes.

Blake shares a personal example that drives home the value of advisory over compliance. His tax preparer advised setting up a C-Corp to potentially qualify for QSBS treatment, which could save millions in taxes someday.

“I can’t even quantify the value of that,” Blake says. “But that’s why I’m willing to pay thousands of dollars for a tax return. It’s that insight, not the return.”

Meanwhile, DIY tax software keeps getting better. Blake describes doing a business return himself using consumer software with ChatGPT open for research. The same process would have taken hours of manual work just a few years ago.

Firms that stick to just preparing returns are going to get squeezed from both ends.

“AI-enabled professionals and firms, they’re going to outcompete and outperform,” Elizabeth warns, “because they’re going to be able to do it faster, better and get to this advisory, which our clients want.”

What to Do Right Now

So where should a traditional tax firm start? Elizabeth recommends figuring out what you hate doing.

“What are your pain points that you hate to do?” she asks. “There’s a pretty high likelihood that I or a talented person on my team is going to be able to say, ‘This is how we can solve that for you.’”

The technology exists today. SurePrep can handle document gathering. SafeSend can automate delivery. Ready to Advise can help you identify tax-saving opportunities. CoCounsel can answer complex questions using curated, expert-verified content. The audit logs are there to verify everything the AI does.

The harder change is mental: accepting that the compliance work that defined the profession for decades is becoming commoditized, and the future belongs to firms that embrace automation as the foundation for higher-value advisory services.

Elizabeth even suggests bringing these concepts into accounting education to attract new talent. Currently, tax courses focus on rules and calculations rather than strategy. After all, accounting is still “the language of business,” as Elizabeth was told as an undergraduate. The difference is that AI can now handle the grammar and spelling, freeing professionals to focus on telling the story.

The transformation won’t be easy, but it’s not optional. As Elizabeth learned when she left the profession out of frustration with mundane tasks, talented people won’t stick around if the work doesn’t engage them. The good news is that automation finally makes it possible to eliminate the drudgery and focus on what really matters: helping clients succeed.

Listen to the full conversation with Blake, Elizabeth, and Kirat for more insights on preparing your firm for the automated future of tax.

From OnlyFans Audits to AI Cheating Scandals: Inside Accounting’s Strangest Week Ever

· ·

In episode 465 of The Accounting Podcast, hosts Blake Oliver and David Leary tackle one of the most bizarre unintended consequences of recent tax legislation: IRS agents may soon need to review OnlyFans content at work to determine if digital creators qualify for tax deductions. This absurd scenario perfectly captures the chaos unfolding as artificial intelligence and new regulations collide with traditional accounting practices.

The IRS’s Awkward New Job Requirement

The new “no tax on tips” deduction allows digital content creators to deduct up to $25,000 from their taxes. But conservative groups successfully lobbied to exclude “pornographic activity” from this benefit, leaving the IRS to determine what qualifies as pornography—a definition the Supreme Court has never clearly established.

“Are IRS agents going to have to sit in their offices at work and look at OnlyFans accounts and determine whether or not this content qualifies?” Blake asks. “Supreme Court Justice Potter Stewart famously said, ‘I know it when I see it.’ So that’s my question.”

The timing couldn’t be worse. The IRS just closed hardship telework requests, forcing employees back to the office while the agency faces a backlog of over 8,000 accommodation requests and has lost 25% of its workforce through voluntary separations this year.

David raises another complication: “If somebody did one video that got determined to be pornographic, do you lose the whole deduction or can you claim all the other days that you got tips?”

Tax professionals face their own dilemma. “Let’s say you get a client who says they want to claim the tips deduction, and they’re an online creator,” Blake explains. “Are you going to check out the content and decide whether it qualifies?”

When AI Meets Ethics—The KPMG Scandal

While the IRS grapples with content moderation, KPMG Australia is dealing with its own technology-related embarrassment. Multiple auditors were caught using AI and group chats to cheat on mandatory compliance training during 2023-2024. This happened after KPMG had already paid a $50 million fine for exam cheating from 2015-2020.

“AI is really good at taking these kinds of tests,” Blake notes. “Just copy paste all the questions into ChatGPT and you’ll pass in a heartbeat.”

The consequences were light: formal warnings for most, one verbal caution, and one person who left months later. The firm didn’t report the incident to regulators.

“They got fined $50 million for it before and then they just continued to do it,” David points out. “So the fines don’t work, obviously.”

The PCAOB is now warning it will closely scrutinize AI use in accounting firms. They’re particularly concerned about private equity-backed firms, fearing pressure for short-term results will compromise audit quality when combined with AI automation.

The Death of the Billable Hour

Beyond scandals, AI is reshaping how accounting firms operate and charge for their services. The billable hour, introduced in the early 1900s as a management tool and dominant since the 1960s, faces extinction.

“When AI can review thousands of contracts in minutes instead of weeks, charging for time spent becomes economically absurd,” writes Rita McGrath of Columbia Business School in the Wall Street Journal.

Blake experienced this transformation firsthand as a freelance bookkeeper. “I billed hourly for keying transactions into accounting software. I then figured out how to automate 90% of it. I had a choice: bill 80-90% fewer hours and lose all my revenue, or switch my clients to fixed fees and take ownership of the process.”

The efficiency gains are already here. Ramp has AI approvals handling 80-90% of transactions automatically. Xero’s new auto-reconcile feature uses AI to match transactions with high confidence. According to OpenAI’s survey of 9,000 workers, employees save an average of one hour daily using AI, with heavy users saving ten hours weekly.

But not every company succeeds at this transition. Pilot raised $118 million at a $1.2 billion valuation, betting it could automate bookkeeping and achieve software margins. Today, they have just 2,500 clients and recently launched a partner program to offload the labor they couldn’t eliminate.

“The fact that they’ve launched a partner program indicates they’re trying to push labor costs out of the company so they can be a software company,” Blake observes.

The irony isn’t lost on David. “They have this headline, ‘Tired of endless QuickBooks updates breaking your workflow.’ But the very first app they list in their integrations is QuickBooks. It’s built on QuickBooks.”

AI Writing Reports Nobody Trusts

Companies are racing to use AI for financial reporting even while harboring deep doubts about its reliability. Twenty-eight percent of financial executives already use generative AI for external reporting. ON Semiconductor’s AI writes entire sections of management discussion and analysis. Hewlett Packard Enterprise plans to use AI for first drafts of financial statements starting in January.

“Take financial statements, drop them into ChatGPT and ask for the narrative. It does a spectacular job,” Blake says. “Taking numbers and turning them into a story that non-accountants can understand, highlighting what’s important, it’s really good at that.”

Yet Harvard Business Review’s survey of 603 business leaders shows only 6% of companies trust AI for core business processes. Most limit AI to low-risk or supervised tasks.

“The work accountants do requires near 100% accuracy,” Blake explains. “Research shows AI achieves 80% accuracy at 30-minute tasks but 100% only for tasks taking a few minutes.”

Meanwhile, Meta’s creative accounting for its Hyperion data center—using complex structures to keep it off-balance sheet—shows human financial engineering still outpaces AI. As the Wall Street Journal called it, “Artificial intelligence, meet artificial accounting.”

What Comes Next

Interesting research is challenging assumptions about what drives audit quality. Studies show offices with less competition deliver better audits with fewer errors. “Competition pushes down fees, which incentivizes auditors to cut corners,” Blake explains.

Another study found audit teams with more women deliver higher quality at lower fees, but only in supportive environments with good work-life balance and female partners.

President Trump, meanwhile, claims tariff revenue will eliminate income tax entirely. “We’ve taken in literally trillions of dollars,” he stated, though actual tariff revenue was only $258 billion last year versus $2.7 trillion from income taxes.

“Doesn’t anybody prep him?” David wonders. “He just makes up numbers.”

The accounting profession is at a crossroads. Will accountants become the quality control layer ensuring AI meets professional standards? Or will they cling to outdated models until technology makes them irrelevant?

To hear Blake and David’s full discussion, including details about the new Trump IRA accounts for kids and Senator Jim Justice’s $5 million tax settlement, listen to episode 465 of The Accounting Podcast.

Your Excel Data Never Leaves Your Computer With This AI Automation Method

· ·

While 58% of professionals have tried AI, only 17% use it regularly. Kyle Ashcraft sees opportunity in that gap.

In episode 108 of the Earmark Podcast, host Blake Oliver sits down with Kyle, a CPA who built Maxwell CPA Review and helped over 1,500 students pass their exams, for a live demonstration that might change how you think about Excel automation. Their conversation shows how any accounting professional can start automating their work in under an hour. No coding experience required.

The AI Gap Nobody’s Talking About

“The more advanced AI becomes, we can take one of two directions,” Kyle explains during the demonstration. “You can continually veer away from it, and the more that comes out, you step farther and farther away from it. Or you can make it a goal to learn, let’s say, one new tool a week.”

The problem isn’t that accountants don’t want to use AI. It’s that they don’t have dependable strategies for implementing it. Kyle describes the typical approach as, “Opening up ChatGPT, throwing in a spreadsheet, and then giving it a prompt and seeing what it comes up with. Sometimes like a Hail Mary, where you just want to see if it gives you an acceptable output.”

There are two major issues with this approach. First, it often takes multiple attempts to get the output you want because ChatGPT can’t read your mind. Second, and this is crucial for accountants, when you upload a spreadsheet to ChatGPT, “your Excel document is going directly to OpenAI. Your prompt is going to them, and the prompt that they output to you is going to them as well.”

This matters because OpenAI’s data retention practices are questionable at best. They’re currently in a lawsuit with The New York Times and required to permanently retain logs. No wonder 70% of accounting professionals cite data security as their primary concern with AI adoption.

Enter “Vibe Coding”: When Everyone Becomes a Developer

Kyle’s journey started with a challenge. Could someone with zero coding experience build something that traditionally required a development team?

Four months later, he had his answer. Using Cursor, ChatGPT, and Claude, he built a complete assessment platform that identifies students’ weakest areas, emails follow-up practice materials, and provides analytics dashboards for professors. All with no programming background whatsoever.

“This really shows it’s possible to not have any idea what the code itself is saying, but with clear communication and patience, you can accomplish things that would have been impossible just two years ago,” Kyle tells Blake.

This phenomenon has a name: vibe coding. It’s coding without being a coder, using everyday language to generate complex scripts. During the demonstration, Kyle shows how Cursor generates hundreds of lines of Python code based on simple English instructions. You don’t need to understand what those lines mean, you just need to know what you want to accomplish.

Kyle offers a metaphor that reframes the entire relationship with AI. “Picture it like an orchestra and a conductor. You’re the conductor. You are in control. You set the tempo. You set the vision of what you want to achieve. And it’s the orchestra that’s doing all of the hard work.”

“There’s this assumption that AI is going to eliminate a lot of work,” Blake observes. “But what we find in reality is that it shifts the work from doing to reviewing. So that job is not going away, but now we review the output and provide feedback.”

The Script Solution: Privacy and Reliability in One Package

During a live Q&A, one attendee asks the question on everyone’s mind: “When you load the project into Cursor and it shows you the Excel files, does this AI platform not retain that client data? How is this different than uploading the Excel into ChatGPT?”

Kyle’s answer reveals why scripts are game-changing for accounting work. “It does not retain this data because with this process, it created this Python script, which is just Python code. It’s offline. There’s no record of this script.”

Your Excel data never leaves your computer. Instead, AI creates a script—basically a recipe—that runs locally on your machine. Think of it this way: instead of handing your sensitive client data directly to an AI company, you’re asking AI to write you instructions. The AI writes the instructions based on your request, but it never sees your actual data.

Blake highlights another advantage: “When Cursor communicates with AI services like Claude, it does so through APIs that have zero data retention policies. That’s in stark contrast to the chat interfaces most people use.” As he explains, these companies want large enterprises to be comfortable, so API interactions have much stricter privacy protections.

But privacy is only half the equation. Scripts also solve the reliability problem. Blake shares a cautionary tale about a Big Four firm in Australia that had to refund a government contract because its AI-invented citations didn’t exist. “They send an entire report to the government, the government clicks on the links for it, and they don’t exist. It’s disastrous if you don’t actually review the output.”

When another attendee asks about the risk of hallucinations, Kyle explains why scripts are different: “You’re not having an AI model interact with the Excel information. You’re having this step-by-step script that says, ‘do an auto sum of column B.” The script uses Excel’s own functions, it just automates the clicking and typing you’d normally do manually.

This deterministic nature means the same script produces the same result every time. As Blake notes, “We can reuse the script we created, apply it to a new Excel file and get the same expected result without having to check everything over again.”

The Three-Part Formula That Makes It All Work

“Goal. Steps. Output.” With these three words, Kyle unlocks the secret to making AI do exactly what you want.

During the demonstration, he tackles three real-world Excel challenges that every accountant faces. First up: a messy data export with empty rows, headers in row three, 14 different date formats, and inconsistent spacing.

His prompt is elegantly simple:

  • Goal: Clean up this Excel file
  • Steps: Identify any inconsistent formatting. Add basic color and style. Analyze each column to better understand its format
  • Output: A new Excel document

Within moments, Cursor generates hundreds of lines of code. The result is a perfectly formatted table with consistent dates, proper headers, and professional styling. “It looks clean, smooth, with some nice shading,” Kyle observes. “It’s just easier to look at overall.”

When Blake asks whether Cursor can do its own checksum, they quickly add both files and ask Cursor to verify nothing was lost. The response: “All 20 transactions are present. All amounts were correctly processed. The sum of $19,000 is maintained.”

The second demonstration scales up the complexity. Kyle shows a General Ledger detail export with 400 rows spanning every account. Manually organizing this would require hours of filtering and copying. His structured prompt creates a summary tab showing account codes, transaction counts, debits, credits, and net amounts, plus individual tabs for each account’s detailed transactions.

“Instead of going to each account in your accounting system and exporting the GL individually, just export all the accounts together and then run this through,” Kyle suggests. What might take an hour completes in under a minute.

The third example addresses bank reconciliation, comparing statements to GL detail to find discrepancies. No more scrolling row by row. The automation identifies matching items, missing transactions, and differences between the files instantly.

Blake connects the dots for viewers. “I picture our listeners who work with some older ERP systems that don’t have very customizable reporting and who are doing a lot of manual formatting. Now you can automate that recurring task every month or every week.”

Getting Started Is Simpler Than You Think

The transformation begins with two downloads that take five minutes each. First, download Python, then download Cursor. Start with the free tier. Kyle uses the $20 monthly plan for daily use, but the free version is powerful enough to begin.

When you first open Cursor, it will ask you to install some packages like “pandas” for Excel interaction. Kyle recommends, “Click the dropdown button and choose ‘run everything’ so you trust the platform. It’s very reliable, and then anytime it needs a new required package, it automatically downloads that.”

Don’t forget to adjust your privacy settings. In Cursor’s settings menu, scroll to privacy options and select “privacy mode” with “no training data used.” This ensures your work isn’t incorporated into AI training datasets.

The key to success is to start small and be patient. “Try it with some information that is not private at all, maybe one of your own documents,” Kyle suggests. “The more patience I have, the more I follow up on that review step by giving it tiny pieces of feedback, the more it improves over time.”

Blake adds perspective on managing expectations: “When I try new tech, 80% of what I do doesn’t have a payoff, but then the 20% has a huge payoff. So don’t get discouraged if your first few attempts fail.”

For recurring tasks, the payoff compounds quickly. “Private roles always have month-end closing. Public end clients always need amortization and depreciation schedules for their notes,” Kyle notes. Even creating client checklists based on prior year information becomes a candidate for automation.

The Bottom Line: Your Move

The tools are accessible. The knowledge is available. As Kyle demonstrated with live examples, you can go from messy data to polished reports in minutes using nothing more than clear English instructions.

So, will you step away from AI as it advances, or learn one new tool at a time and stay connected to this movement? Because as Kyle reminds us, “It’s not going to go away. It’s just going to become more integrated into everyday work culture.”

To hear these demonstrations in action, listen to the full episode at podcast.earmarkcpe.com/108. Kyle has also offered to help early adopters, so reach out to him at kyle@maxwellcpa.com with questions or to brainstorm how this could apply to your specific work situation.

As Kyle challenges at the session’s close, “Try your first task with it this week and see how it works for you.” The revolution in accounting work is here, waiting for those bold enough to embrace it.