As the world continues to digest the rise of generative AI, agentic AI lies waiting on the bleeding edge, and while few accounting firms are using it at the moment, major players in the space have already made significant investments in what they believe to be the next step in the AI revolution.
Very broadly, an AI agent is software that is capable of at least some degree of autonomy to make decisions and interact with things outside itself in order to achieve some sort of goal—whether booking a flight, sending a bill or buying a gift—without needing constant human guidance.
The concept of an AI agent is not new, as computer scientists and software engineers have been using the term for years, and such agents are already used in commercial applications — Sage Copilot, for instance, uses purpose-built AI agents each with their own area of specialization whose efforts are coordinated Copilot, which acts as an interpreter between the AIs and the human users who are requesting they perform a task.
AI using AI
Antony Weerut – stock.adobe.com
Given this one might wonder why interest in agentic AI and AI agents seems to have risen only towards the end of last year (at least as measured by the volume of Google searches on the subject). One answer is that while agents have been used for years, advances in generative AI made them much easier to create and deploy, according to Hamid Vakilzadeh, an accounting professor at University of Wisconsin at Whitewater who has written extensively about AI.
“Agents have been around but we had to program them in a logical format. But because of large language models who can understand natural language, it makes it much more flexible to create very sophisticated systems without having to know so much coding, and it’s much easier to implement on a larger scale,” he said.
He contrasted this with classic AI models.
“If you look at a machine learning thing like recommendations, those are pretty sophisticated, they’re pretty useful in today’s market in entertainment but those’re not really doing a task, they make up a menu like if you need to find Christmas movies. They don’t make a decision by themselves, you make the final decision that I am going to see this, they propose but you make the decision. [In contrast] an AI agent can accomplish a task,” he said.
Beyond this, advances in generative AI have also made AI agents themselves more effective in the field. Pascal Finette, co-founder and “chief heretic” at tech advisory firm Be Radical, contrasted this with robotic process automation. While he said RPA is not to be underestimated, even today, it tends to be very rigid in its setup, operating mostly on if/then/else principles that work very well for defined use cases but struggles in the face of unstructured data or unusual edge cases. Agents, bolstered with generative AI, become much more flexible.
“The reason I think why this is happening is we now have this superpower of an LLM which allows us to look at the world and look at data in a much more unstructured way and still get some really interest insight from it which we can then use to automate stuff, to execute on our behalf. … the beauty of LLMs and gen AI is it has the flexibility to be able to actually create meaningful interactions,” he said.
David Wood, an accounting professor at Brigham Young University whose research also heavily involves AI, noted that ‘agents’ can be thought of as a framework for applying technology; agents are programmed to do a task, and they can use other tools to accomplish that task, and so rather than being some sort of evolution from traditional RPA or generative AI, an agent can be thought of as something that will use RPA and generative AI.
“This is a different framework for how we do programming. We program an agent to do something, it could be to use generative AI, it could be to do a machine learning algorithm, it could be to simply change the color of the font. You can program an agent to do what you want and agents can work together or even compete against each other to do something, so it is not just a generative AI topic but highly valuable now because agents can use generative AI,” he said.
This increased flexibility has led to major investments in the technology from significant players. Big Four firm KPMG, for example, announced in October a minority equity investment in Ema, an agentic AI startup building universal AI employees as part of the firm’s overall vision of action-oriented assistants working seamlessly alongside and augment human teams.
Around the same time, accounting solutions provider Thomson Reuters announced it had acquired Materia, a U.S.-based startup that specializes in the development of an agentic AI assistant for the tax, audit and accounting profession. This transaction, which is complementary to Thomson Reuters AI roadmap, accelerates Thomson Reuters vision for the provision of generative AI tools to the professions it serves.
That same month, Microsoft announced the addition of its own agentic AI capacities, namely the ability for users to create their own autonomous AI agents with Copilot Studio as well as the release of ready-made in Dynamics 365 that can handle things such as sales, finance and supply chain management.
Despite these high profile announcements, though, the field is very young, with many applications still in the experimental phase. Finette said it isn’t even necessarily bleeding edge so much as jagged edge. However, based on announcements like these, it appears this is the direction the AI community wants to go next.
Wood agreed, saying there are not a lot of agentic AI solutions right now that are fully production ready, but he sees great potential in the technology once it grows to maturity. For example, many accounting firms bill on how time is spent, which can be very time consuming to effectively track. Agentic AI would be able to observe an accountant work on company A for 45 minutes and company B for 60 minutes and bill accordingly. He said this might lead to people getting rid of all timekeeping because a computer can do it for them.
He also raised the idea that it could greatly increase efficiency for audits. Imagine, he said, if an agentic AI bot could automatically do most audit confirmations, send them to the humans for approval, and flag the things it couldn’t do itself, “so you could build tools for end to end processes to do full tasks together.”
Finette also saw great potential, saying it could act like a full AI worker capable of complex tasks. He said people eventually should be able to go to their AI and say they’re having a meeting in two days with someone and they need a flight and a hotel within their preferred parameters (e.g. cost, distance, etc.) The AI would then perform all the research, compare prices, maybe even generate its own spreadsheet to aggregate all the options, then make a judgment call on which flight to book and which hotel to reserve and actually do it. While an agent might struggle with novel tasks, for the most part it should be able to handle most of the routine work.
“You can translate that into a tax practice, where you have these complex workflows which a human breaks down into individual steps, each influencing the next: you take a document, extract the info from the document, put it into your accounting system, classify it, and do the booking in the system. All of this in theory agentic AI should be able to do for you,” said Finette.
Other accounting-specific applications he could envision include anything that has to do with data entry, reconciliation of accounts and classification of information in systems, as well as expense management, which he said is already semi-automated already.
“Right now it is semi-automated where you upload something into Expensify or something and it does image recognition on those expenses and pulls them in, but in the future it should do the report for me, there’s no reason why it should not take all this information and put the report together and submit it on my behalf,” he said.
However, Wood warned that agentic AI will still carry many of the risks that generative AI has today, especially the risk of making up information or being inconsistent with its outputs. While companies might promise a genie-like wish fulfillment, it will be especially important for people to understand the limitations of this technology.
“These systems interact, you wont always get a deterministic outcome like they think computers will generate, its not a calculator, so if you give an agent the ability to be creative, sometimes it might produce output A and sometimes produce output B and in accounting and business that can be a great strength, like in marketing, but when you do a tax form you don’t want that, you want income to be correct every single time. So the risk is that everyone gets hyped up and excited and applies it in the wrong place, you gotta use these tools and understand their strengths and weaknesses,” he said. “It’s sort of like gen AI right now, they think it will solve everything, but it solves these specific sets of issues and problems, so knowing where to use it and how to use it will be important.”
Finette agreed that the tendency of generative AI to make up information would still be a risk, and that there will likely be a lot of hype trying to minimize this risk as well. But he also noted that the fact that agents can actually act semi-autonomously and make decisions means the consequences of these risks can be bigger.
“In the flight booking use case, do you really trust the AI to actually book the flight for you? Will you be on the right flight at the right time? This is a silly example but a very real one,” he said. “The other [risk] is when you let AI make ‘moral’ decisions like letting AI do promotion decisions or suggesting out of 100 people here are the people who are top performers, where you get into issues like bias, which we know exists in AI. So all the issues we have with AI will be amplified with agentic AI.”
While theoretically these AI agents will be supervised by humans, Finette wondered about the degree to which people will actually do so, especially when AI can be so convincing in its reasoning even when wrong.
“These systems are so overly confident in their responses it is hard for some humans to step back and say don’t trust it. We all have experiences where you use ChatGPT and it tells you something wrong but they tell it to you in such a convincing way that if you didn’t have the knowledge you’d take it as gospel. … It is amplified if you let the system execute on this information. The human challenge is, and there are a bunch of research papers showing AIs are as convincing or even more so than humans, we need to get our workforce to understand that they should tread with caution and not let the AI bully you into a corner,” he said.
Billy Long speaking at a Donald Trump campaign event
Al Drago/Bloomberg
The week before confirmation hearings for President Donald Trump’s nominee for commissioner of the Internal Revenue Service, former Missouri Congressman Billy Long, Democrats in the Senate are asking questions about the timing of campaign donations he received immediately after his nomination.
In a letter sent last Thursday to seven different companies — including an accounting firm, a tax advisory services firm, and a financial services provider — Democratic Senators Elizabeth Warren, D-Massachusetts, Ron Wyden, D-Oregon, and Sheldon Whitehouse, D-Rhode Island, questioned donations that the companies and some of their employees made to Long in the month and a half after his nomination in early December of 2024.
Between Dec. 4, 2024, and the end of January 2025, the letters said, Long’s unsuccessful 2022 campaign for Senate received $165,000 in donations — after nearly two years without receiving any — and his leadership PAC received an additional $45,000.
The donations allowed Long to repay himself the $130,000 balance of a $250,000 loan he had personally made to his campaign back in 2022.
The senators’ letters described the donations as “a highly unusual and almost immediate windfall,” and characterized many of the donors as being “involved in an allegedly fraudulent tax credit scheme.”
“The overlap between potential targets of IRS investigations and the list of recent donors heightens the potential for conflicts of interest and suggests that contributors to Mr. Long’s campaign may be seeking his help to undermine or avoid IRS scrutiny,” the letters said; adding, “This brazen attempt to curry favor with Mr. Long is not only unethical — it may also be illegal.”
The senators then warned, “There appears to be no legitimate rationale for these contributions to a long-defunct campaign other than to purchase Mr. Long’s goodwill should he be confirmed as the IRS commissioner,” before appending a list of approximately a dozen questions for the donors to answer.
The donations were originally discovered in early April by investigative news outlet The Lever, which the senators noted in their letters.
After Long left Congress in 2023, he worked for a tax consulting firm, including promoting the COVID-related Employee Retention Credit. In early January, Sen. Warren sent a letter to Long questioning his tax credentials and promotion of the ERC.
President Donald Trump called on members of his party to unite behind his economic agenda in Congress, putting pressure on factions of lawmakers who are calling for last-minute changes to the legislation to drop their demands.
“We don’t need ‘GRANDSTANDERS’ in the Republican Party,” Trump said in a social media post on Friday. “STOP TALKING, AND GET IT DONE! It is time to fix the MESS that Biden and the Democrats gave us. Thank you for your attention to this matter!”
Trump sent the post from Air Force One after departing the Middle East as the House Budget Committee was meeting to approve the legislation, one of the final steps before the bill can move to the House floor for a vote.
House Speaker Mike Johnson has set a goal to pass the bill next week before the House recesses for its Memorial Day break.
However, the the bill failed the initial committee vote — typically a routine, procedural step — with members of the party still sparring over the scope of the cuts to Medicaid benefits and how much to raise the limit on the state and local tax deduction.
Narrow majorities in the House mean that a small group of Republicans can block the bill. Factions pushing for steeper Medicaid cuts and for an increase to the SALT write-off have both threatened to defeat the bill unless their demands are met.
“No one group gets to decide all this stuff in either direction,” Representative Chip Roy, an ultraconservative Texas Republican advocating for bigger spending cuts, said in a brief interview on Friday. “There are key issues that we think have this budget falling short.”
Trump’s social media muscle and calls to lawmakers have previously been crucial to advancing his priorities and come as competing constituencies have threatened to tank the measure.
But shortly after Trump’s Friday post, Roy and fellow hardliner Ralph Norman of South Carolina appeared unmoved — at least for the moment. Both men urged continued negotiations and significant changes to the bill that could in turn jeopardize support among moderates.
“I’m a hard no until we get this ironed out,” Norman said. “I think we can. We’ve made progress but it just takes time”
While CPA firms far and wide have made major technology investments over the years, the vast majority of accountants say they’re not being used to their full potential.
This finding comes from a recent survey undertaken by CPA.com and payment solutions provider Bill. The 400-person poll found that nearly all respondents, 97%, say they use technology inefficiently and that additional training is needed to maximize return on investment. Further illustrating the point, 43% of respondents said that technology is making them do more manual work, not less, something. Becky Munson, an Eisner Amper partner specializing in outsourced accounting services, believes this reflects a failure of training and change management, as she has seen many who disliked a technology change develop manual workarounds specifically to avoid using the new solutions.
“We see employees make workarounds with tech stacks, which makes headaches that I think align with this 43%. We train people on new things, we ask them to use them, and they keep doing what they were doing before and only use the technology as much as they have to [in order to] move things along while you have people well trained on the software keeping up,” she said in a webcast on Thursday about the survey.
Ariege Misherghi—senior vice president and general manager of accounts payable, accounts receivable and the accountant channel—said the issue isn’t just because of firms but also vendors that don’t provide enough support, and may not necessarily understand the profession in the first place.
“Too often I think tools aren’t fully aligned with the workflows they’re meant to support. In SaaS they talk about product-market fit, but in this profession it’s not just that but also product-firm fit, and maybe product-profession fit. Not every tool marketed to accountants was built by people who truly understand how this profession works: the rhythms, the regulations, the stakes, the relationships, all of that. And even the greatest tools can fall short if they’re not implemented with a deep understanding of how firms really operate,” she said.
And sometimes the inefficiencies come from both sides at once: the survey found that only 37% of firms require clients to use their tech stack, something that Munson said “breaks my heart” as “it is so low.” A streamlined, established tech stack is needed to achieve true economies of scale, but to get there firms need to standardize their data, and to do that firms need to make sure their clients’ data is also standardized, which usually means integrated tech stacks.
“If you have all these different clients with all these different technologies, even if your own tech stack is standardized the systems they use is different, so the kind of data you will get will be different, and the work you need to do to make it work with your data is different, and your team spends a lot of time spinning their wheels,” she said. “Once you get standardized, where everything back and forth from clients is the same, you get to see how well the teams can do their work.”
One source of inefficiencies is a rushed implementation. Munson said that, too many times, firms are so eager to get a solution working that they don’t pay attention to all its capacities, just the ones they need right now, but once the basics are down firms still don’t circle back on the rest of the features and how they can be used to drive efficiency.
“Most of us have been through an implementation, either in the practice or with a client, where you’re just like ‘anything to get it working. Forget about all the fancy things it does. We just needed to do the basics right,’ and then we never circle back on those better, more efficient processes. We get to sort of minimal viable, and then we forget to come back and give it an extra polish. And so what we see there is the processes get written for that basic piece, and we never update,” she said.
But this is part of what both speakers believed was the larger problem of firms getting lost in the details of their tech stacks and not taking a broader, more holistic approach, which would enable more efficiencies. The key component to managing technology effectively, Munson said, is looking not at individual solutions here and there but thinking of the system as a whole.
“Often, what happens is something’s wrong or something is troublesome in some way. And so [we say] what can we do to fix that one thing? And we don’t think about it holistically and get all the right folks in there so that we’re solving for the right pain points,” she said.
Misherghi agreed, and added that this holistic extends not only to the technology a firm already has but the solutions they plan to purchase in the future. When evaluating what technology they need, she said leaders need to think not in terms of specific point solutions to particular problems but things that can support the entire workflow—plus, the onboarding, training and ongoing support from the vendor.
“Don’t just look for features, right? Look for solutions that support your workflows from providers that understand you. For firms, onboarding and training and optimization can’t be an afterthought. They’re essential to realizing value. I think this is where vendor partnerships matter. Firms seeking the strongest results aren’t just using software, they’re collaborating with their providers, they’re staying educated, they’re making sure their tools evolve alongside their needs. The best outcomes happen when your technology partner acts like part of your team, not just part of your toolkit,” she said.
Misherghi said that the more successful firms she’s seen think less in terms of performing particular tasks but designing an entire system that, through automation, can do those tasks for them. It is less about plugging holes and more about developing a full infrastructure. The survey found that 74% of participants have a detailed plan to add new services in the next 12 month; Misherghi noted that, among these firms, 86% have a detailed technology roadmap, which is “a wonderful mark on the evolution of the profession we’re seeing.”
She said a good tech roadmap is more like a service design blueprint versus a shopping list. Successful firms, she said, are not just chasing features but designing intentional workflows and systems capable of scalable service delivery. Similarly, she stressed that the provider should be more than just a vendor but a strategic co-architect that can help with growing pains.
Misherghi said this approach will become especially relevant as AI becomes more common, as integrations will be key to their effective use, which means thinking in terms of the whole system to understand where those integrations should take place. Right now, she said, people think of AI in terms of analyzing data or extracting fields, but with the rise of AI agents will require firms to focus more on coordinating between them.
“I think the next big leap is when those systems don’t just talk to each other, they act on each other’s behalf. I think the next big inflection point will be moving from automated steps to autonomous workflows, where AI agents aren’t just analyzing data or extracting fields but actually orchestrating tasks across tools based on firm policies and context and that will change the role of the accounting profession: its less time doing the work and more time designing the system for how everything works together. So the firms that will be thriving are those who are building strong infrastructure now because that is what AI needs to deliver on its core value,” she said.