
Loading summary
A
Welcome to the AICPA Town Hall Series, your resource for the latest news and updates on pressing issues facing the accounting profession.
B
Good afternoon and welcome to the AICPA Town Hall. I'm Eric Ouskerson, one of your hosts for today. Today is April 9, 2026 and I'm here in the New York studio with Mark Koziel. And Mark, we've been at this now for six years because Mark Cozy and I kicked off the AICPA town halls back in 2020. That was April 3rd. And then since then, Mark, we've had 168 shows. That's an average about 28 a year. Because when we started, Mark, remember that we were doing them every, every week. So we're here to talk about the most pressing issues of the day. It's just a fantastic way to connect with all the firms and members in business and industry.
C
Absolutely.
D
I mean, just to come out of during a pandemic to bring us all together to create this community and for us to provide information. And it was you and I, 15 hour days saying we need to talk to our members, we need to tell them what we know and what we don't know. And now to have that venue for our members on a regular basis is fantastic.
B
Yeah, so we kicked that off in April of 2020. We were all at home. Today we're in the studio. We had a couple thousand attendees back in 2020. It's continued to grow and we're averaging 15,000, close to 15,000 live attendees. So today, Mark, we got a, we got a great program. We're going to kick things off with a quarterly profession update from you. Then the DC team, Rachel Dresden is going to give us a DC update. I'm excited about this tech technology leadership discussion that we're going to be having with Pascal Fannette. A lot has changed related to the AI models. We'll be diving into that. Then we've got a technical update with Daniel Huff. And then, Mark, I know there's another important topic related to AI and koso.
D
Absolutely. And it's, you know, when we think about what our opportunities are going forward as a profession, when we talk about AI and controls over AI, it's going to be a good topic.
B
Okay, so let's, let's, let's get right into it. A recent AICPA Harris poll.
D
Yeah, you know, I had a question by team when I got this information and there was about to be a press release to say that, you know, there's a Harris poll that says that America, Americans who are Delaying major decisions for financial reasons is actually down from where it was in 2025. 61% in 2025, 55% now in the in the most recent poll, Americans say that they feel better about their financial situation over the next 12 months. 36% up slightly from the 33% we were in 2025. Now in 2025, this is pre HR1 bill and a variety of other things that were happening. But you know, the sense is and people immediately when I saw this, I'm like, yeah, but this is before Iran. And they're like, no, it's not.
B
And I'm like, 28th is the start date.
D
Yeah. So, you know, understanding my dates. And so you had we were already in Iran. Definitely far more uncertainty on it then than now. And then also the DHS shutdown. And what was going to happen with tsa. Am I going to hold off on traveling because of what's happening in TSA lines? And I think we're generally, what we're starting to see is the economy just keeps kind of pushing through it. You had mentioned this before, the recession that never happened last year. We continue on, Americans continue to spend money and we'll see if that continues. But right now all indicators are that we're generally we're no worse off than where we were last year. And the economy's been pretty resilient over the last couple years.
B
Well, the tech investment, we're going to be talking about that tremendous. And the US Is leading the way with these AI models.
D
Absolutely. Data centers and chips and a variety of other things that are happening. I saw an article about, you know, production of steel and trying to get ships built again in the US So there are a lot of things, there's a lot of factors that go into that. And we've talked a lot about the tariffs. What was the effect of that going to be? And even with the tariffs, you know, the Americans still spent through the tariffs as all of that uncertainty was unfolding over the last year. So we go to the next slide. You know, I'm going to take it easy on our members who many of them are still in busy season. We got a week of that left and maybe none of them are on today because they're all still getting personal filing done. But you know, there's a couple of big initiatives we started last year and the first was Rise 2040. The profession's done a great job here in the US of kind of visioning out 15 years. Started in the late 90s with the Vision Project, CPA Horizons 2025 that was in 2011. So here we are with 2025 expiring. We've used so many different technology tools to really be able to reach out to our members on a regular basis. Over 6,000 professionals around the glob 25 countries that you see there are 90 plus future forums. So we have a good representation of our members. Small firm, mid sized firm, large firm, business and industry, business and industry in the U.S. business and industry abroad. It really was a good cross section. And so here are some of the top trends. We're finalizing the report as we speak. I have the association board meeting in a week and a half and at that board meeting we are going to get the hopefully board's approval of what the report looks like. One of the big things, no matter where we went around the world are those big three words, trust, ethics and integrity and them still carrying the day. That's what we were built on. That's what we are still about today. And we're going to talk about gen AI and internal controls over AI and why it is such an opportunity for us. When we think about we are still the trusted profession in business that being able to deliver on that, making sure that we maintain that level of trust. AI technology. We've talked about technology over all the years. Back in the day, you know, 15 years ago, 2025 or 2011 when we had this technology was top of mind. Even in the late 90s technology was top of mind. Now we insert the word AI but AI is just the blend of the day. 10, 15 years ago. Eric, you know this full well. CPA.com was all over it. The advent of the cloud and how that's going to change how we do
B
and it did brought up a lot of opportunity.
D
A lot of opportunity, Ton of opportunity. And you think about client accounting and now. Yeah, I'm not just in my local small geography now. I could do it anywhere around the country. Very different environment and also got us ready for the pandemic to be able to make that happen as a profession. We pivoted very quickly around that pandemic and our firm still did very well during that time. So AI technology moving from compliance heavy to to strategic advisory. We're not saying compliance is going away permanently. We're still going to be in the compliance business. But the amount of time or the amount of effort we put into compliance will be reduced and strategic advisory will still carry the day as we've been touting now for a number of years. So we'll continue to See that talent pipeline, workforce sustainability. We'll talk about talent in a minute. But a lot of what I've heard over the last year, you've heard me talk about this on the town hall is the advent of upskilling our professionals, our young professionals, giving them the right skills. How are we going to train them when they're not doing the debits and credits like we all grew up doing back in the day? And I think that's going to be an important aspect. And we'll talk about how we've created an initiative that's already bubbled out of the Rise 2040 feedback. And then the only real, I don't want to call it a negative, but we have been a profession that some will say that we're resistant to change. And this came out in a lot of the, a lot of the feedback sessions we did. I'm bullish on that, though. I think we have changed. Maybe we didn't come along as fast as everyone else, but we still are change agents in the business community and that'll continue. So stay tuned for more at the Engage conference in June. I'm hoping to see you all there. We're going to have the full report released out then and what you can do about it and how you can integrate that into your business. Creating a strategic plan. Whether you're CFO and business business, small firm partner, trying to grow your business, there's going to be different ways that you're going to be able to connect and get that information. If we go on to the mark
B
just on, you're using a lot of technology. I mean, I look at this project compared to the last two.
D
Yeah.
B
And the level of technology that you're leveraging, I even think you're, you're building like a model here that people will be able to leverage.
D
There will be an AI tool. So any of our members will be able to go into the AI tool. They're going to be able to say who they are and say, these are what I see in strategy. And then you're going to get that level set against what others in your marketplace, in your type of firm or type of organization. All of that is going to be leverageable. And I think there's going to be a huge value add to our members, especially those who are in the CFO community, to be able to say, hey, this is what we're seeing. How does this. Let's get to the C suite and talk about what we're getting from our profession. That's going to be very relevant to what we do into the future. So it's really going to be a great tool. We'll have more about that as we come out of engage next year.
B
Very dynamic.
D
Yep, for sure. And then when we talk about talent, everyone says, is AI going to replace talent? You know, I hear it internally. We have people on our team saying, hey, I'm worried about my job. Is AI going to replace me? No, AI is going to enhance what you're able to do. And I see it as the profession that we're going to be able to really own controls over AI, be able to test against those controls. Also to be that business partner, we have access to data, we understand how to manipulate data. We're able to get in there being insightful about data. So imagine being the business partner. So if you're in business and industry, having a finance business partner with the business lead to be able to do it for a small firm going into a client, being able to help them develop the right prompts to figure out what they want to do with their business, that strategic advisory now the AI is going to give us the ability to be that true strategic advisor. We always said that we always felt like we needed to have the answers. The bigger thing is asking the right questions and AI without a doubt, because we've asked the right questions along the way, it is going to give us the ability to do that as we go forward as well. Profession ready initiative, finding out the skills necessary to be successful. As I said, the new associate today, they're not getting the debits and credits. How do we put that into a simulation model? Imagine a new associate sitting down and they're doing a client interview with an AI enabled tool that is able to get them through that and then they could be actually evaluated right on top of it. You missed this. You could have said this. You should have asked this question. And they're going to be able to get that up front. We're going to have to invest more in our younger professional to get them up to speed because we need them at the manager level in a flash compared to what we all did to say we could wait around for two years doing more of the mundane tasks. We're not going to have that ability anymore. And from this profession ready initiative, coming out with the competencies that are necessary, being able to provide some level of assessment and then a tool to be able to teach or learn against that, that's going to be incredibly dynamic and AI driven. We've had a number of firms ask us already. They said if you build that, we will buy it because we don't want to have to build this on our own.
B
Yeah, I mean, I think it's going to be very exciting. If you look at like the, you know, the PC revolution, people came on board and they had to use PCs and, you know, what was that going to do to kind of the workforce? We may look back on this and say the young professional now can be trained faster than ever with these new tools. So it might be right now. We think it's a problem to be seen.
D
Yeah. And we used to say two weeks and then you're done. Go out into the field. Good luck. Now it may be six months, maybe a year, but they're going to be better trained when they get out into the field then. The final item I have in my short time with us today is around the PCPS Top issue survey. For those who don't know my background, I started with AICPA in 2006. I left for five short years, 2020, halfway through 2020 through 2024. And the 24 came back to 1-1-25. PCPS. I was senior manager of PCPS, the Home for small firms, as we used to say. The top issue survey has been around well before. I started in 06, but that is our way as the AIC to get information from our member firms, to find out what is top of concern and then breaking down those issues by firm size so we can better understand, is it all firm issue, Is it a small firm issue, is it a large firm issue? And then providing tools and resources to be able to help our members navigate through that and hopefully reduce the challenges around it. So I highlight this. This is going to be coming out right after April 15, so take a snapshot of the QR code, wait around a week. I'm sure we'll talk about a little bit more on our next town hall post, busy season. But I, I ask all of you because this is really valuable information for us and it helps give you validation of where you're at, what your issues are relative to other firms your size.
B
It absolutely drives a lot of our focus. So thanks, Mark. And you'll be joining a little bit of discussion with Pascal as well.
D
Yep, absolutely.
B
Thank you. So now let's pivot to dc. And Rachel, I understand that the cherry blossoms are at the later stage of their bloom right now, but they're still out.
E
That's right. So if you want to come to D.C. right now and have good weather and see the cherry blossoms, now's the time to do it. However, if you have allergies, I would probably stay away from D.C. but there's a lot going on right now, even though Congress is in recess during their spring recess. So for two weeks, they are back in their districts and in their states having meetings with constituents. But there's still a lot to report on here. So one of the big things is that just a couple of days ago on Tuesday, there was a special election in Georgia. This was to fill the seat for Congressman Marjorie Taylor Greene, who resigned at the beginning of the year. And what this means is that this was a solid Republican district and the person who won is a Republican. But for the speaker of the House, it means that for that very, very small majority that we've talked about here on the town hall before, there is one more Republican. So now Republicans in the House, if it's a party line vote, they can lose two of their members and still pass legislation. So what this means for us and for, you know, legislation as we look forward is that, you know, they have a little bit more of a cushion. It's still, you know, very narrow, but there are issues like what Mark just mentioned, you know, DHS funding. The Department of Homeland Security has still not been funded. So they are shut down right now. It's the longest shutdown in history for a department. We're going on close to eight weeks now. And so that's one of the biggest issues right now that they are working on. And the issue is, you know, they have concerns and, and differences between Republicans and Democrats on funding for ICE and cbp. So the question is, how do they move forward on that? And one of the things that they're talking about is to fund this through budget reconciliation. So they are talking about using the budget reconciliation process to either fund just ICE and CBP or, or all of the Department of Homeland Security. And by doing it this way, they wouldn't need the 60 votes to pass the Senate. So just as a reminder for funding legislation, they need 60 votes in the Senate. However, Republicans only currently hold 53 seats, and so they would need seven Democrats to vote with them, assuming all Republicans would vote together on this. And so far, they have voted on DHS funding a number of times, but only one Democrat has crossed over and voted with Republicans. So to get around that, they are looking at doing it through the budget reconciliation process, which is how they passed HR1 last year. And it has some limitations by doing it this way. But if they do pursue this, something that's important for the profession is they may include things that could impact us directly. They could include things related to taxes, things that they tried to do last year, but they weren't able to do it because of some of the restrictive rules of budget reconciliation. So they may try to rework those as well as, you know, some tax items that they didn't do last year that they want to include this time around. So that's something that we are watching very closely just because, you know, it could have a tremendous impact on the profession, depending on how they move forward on this.
B
Well, Rachel, I know that the budget request just went in and here's some interesting data related to the departments that are very close to the profession.
E
Yeah. So the President submitted his budget request for FY2027 last week. The President always does this at the beginning of the year. And just something that we wanted to note for everyone. Even though the President will submit his budget, it is not necessarily what will be enacted. So after the President does his budget, then Congress then goes through their process of voting on budget resolutions and then the appropriations process. So you can see here the differences from last year, what was proposed in the President's budget versus what was enacted. And so here we have for FY 2027 what is proposed. But it is rare when any of these numbers stay the same for what is enacted. So just wanted to highlight that for everyone.
B
Well, one thing that we continue to talk about is our state of the nation call to action. And this is a segment that we're going to. We're going to do a longer segment in a future town hall on this.
E
Yeah. So we have a couple of calls to action for the town hall community. And we hope that if you haven't done so yet, that you will just, you know, use these QR codes and it will take you less than a minute to reach out to your members of the House and send to support legislation that is very important to the profession. The first one, as you mentioned, is fiscal state of the Nation. And this has to do with the fiscal health and having the Comptroller General give a report to members of Congress on the nation's fiscal health. And then the other one is including accounting in existing K through 12 grants for students. So we encourage you, if you haven't done so yet, to reach out to your members of Congress.
F
It's really easy.
E
And if you have done so, thank you so much. It does make a big impact.
B
Well, thanks, Rachel. I look forward to having you back in open forum. We do have some questions that came in, so hopefully we'll have some Time to take a couple of them. So with that I'd like to bring up our good friend Pascal Fannette, who many of you know, he's one of the leading innovation experts in the US and actually around the world. And we've been very fortunate to have a long standing partnership with Pascal and Pascal's really helping us. I'm going to move the next slide here then welcome you. But he's really helping us demystify AI for the accounting and finance profession. So his bios in the materials and Pascal, it's great to have you with us today.
A
Thank you again for having me back on the town hall. It's always nice to be here.
B
Well, you're helping us with a lot of these resources and once again, CPA.com, we're the technology and business subsidiary of the AICPA and we're all about really empowering the firms and members in business and industry with all of these different transforming digital technologies. We've been on this project for the last few years related to AI. ChatGPT came out in the fall November of 2022. That's three years old. The first couple of years everyone's trying to put their arms around what Genai was and really where we're now is and you're going to hear from Pascal on this because some things have dramatically changed over the last hundred days is helping firms select the AI tools. Make sense of it. We also have this monthly newsletter which I encourage you to subscribe to and all of these links you can leverage to access these materials. But Pascal, what I want to you and I had a great pre town hall discussion talking about the state of these models and this is something that we're really well aware of due to all of the tech companies that we're working with. But the last hundred days has just been dramatic. I mean it was dramatic when ChatGPT came out. But why don't you share what has happened?
A
Yeah, I think the headlines largely are the model races, right? Like this notion that you have anthropic brings out a new version OpenAI, Gemini, Groh, whatever model you want to use. And I think that's interesting context but really the headline isn't for me about who's winning. That is an interesting question you can ask and if you're a betting person that's maybe an interesting area to invest into. But really the headline for me is not who's winning but what are the capabilities which we are unlocking for practitioners. And really what happened in the last maybe 100 days ish is pretty, pretty remarkable. So if you go just back six months ago, like fall of 2025, we still used mostly just chatbots and we had to put pretty strong guardrails around them. And they were useful, they were good at. I think we all have sent one too many emails refined by a chatbot. We have all probably used chatbots to summarize a document, et cetera. And we were all engaged in this notion about how do we become good at prompting. And what they couldn't do is they couldn't access properly files, they couldn't do multi steps, et cetera. What happened with the latest set of releases, regardless which model you're using, roughly around the time frame of the change of the year, is that we now have these really powerful models which can do really interesting things. They can actually become very productive. They can reason through complex problems and they can do multi steps fairly autonomously and very reliably. And what I personally find interesting for the profession is we're now seeing these interesting signals emerging where initially you could use ChatGPT and upload 1040 information and see if it can do something, and more likely than not, it couldn't. And now we're seeing partnerships such as, you know, in the profession anthropic. And I Win just did a big announcement around a partnership where you actually run the tax computation inside of cloth. So you marry the, the models, the probabilistic models from a company like I Win with the. The generative AI capabilities of an anthropic and then suddenly it becomes something very different and it becomes actually really useful. And I encourage everybody who is maybe a little blase about AI, which I totally understand, it's very noisy, to look carefully what has changed in the last couple of months even, because this is a very different world.
B
Yeah, well said, Pascal. If you just. And Mark, I'll bring you in. But if you look, it is a very exciting time. In particular, the tax area. We've talked a lot about tax research and what these large language models have meant to tax research, and it's transforming that category. But I think for the next tax season, Anthropic Cloud, that tool, it's taking all of the inputs now and then, putting it into a tax engine, I think, and this is. We're talking about putting some events on as we close out this this summer and go into the fall next year to talk about the future of the tax engines which are going to be directly linked to these models.
D
So, Mark and Pascal, forget about the whole idea of what the AI can do, it's how it's going to change and disrupt the business for, let's say, our tax practitioners. I'm already seeing in social media posts that a tax practitioner's client, they sent them their tax return. The client put that tax return into Claude, ran it again and said, hey, why didn't you catch these deductions? And so this whole compliance piece that we had before, they're going to question against it. And now all of a sudden that strategic advisor conversation gets different and should that advisor, should the firm have done that first to say if I were to run this through a AI engine so that I understand what my client may see, I'll be better prepared to have that answer?
A
Yeah, absolutely.
B
It's a good segue to this next slide. So what is happening? Well, Mark, one thing I'll say just on that comment is you could have done that last year, two years ago, sure. The client could have gone to TurboTax, guess what? And that's an AI based model. So it's just, it's just maybe it's happening faster. But this is the clients of the firms, business models of the firms, they want to outsource it to someone and be that advisor. So it's like, I'm not over. It's almost like a new toy out in the marketplace. So Pascal, this is important here, kind of how you, how you break this down and where you think firms should be focusing.
A
Yeah, absolutely. Because with great powers comes great responsibility. And I think one of the more common mistakes I currently see people do is because these models have become so good at writing code, they're actually excellent at writing code these days that we conflate the question of can we build it? With a question of should we build it? And my recommendation for anyone listening here is unless you are a. And we'll talk a little bit more about this later, but unless you are a very large firm with lots and lots of resources, then your equation looks very different. Really think about where do you actually want to invest your time and your energy and leverage these tools to the best of their control capabilities. And we know for a fact that all of the vendors which are currently using and all of the vendors you might be talking to, they're all working on integrating AI in really smart ways into their systems of record. So please, do me a favor, don't come up with the grandiose idea of replacing your GL with a wipe coded, you know, an AI generated version of it though. This is the interesting thing is because these Models have become so good at writing code, they're exceptionally good at writing these little quality of life tools which you can use internally. So stuff you don't expose on the Internet, stuff you don't expose to customer data, but things which can allow you to do things like you've got a large data ingestion, you've got a whole bunch of whatever Excel spreadsheets which need to be converted into something else. And this would have been a manual task for you to, you can actually ask AI to essentially write your code to do this for you. And you know, in my personal practice, we do this all the time. Now, every time I do something three times in a row, I'm like, well, AI should maybe should be able to do this. So my invitation for you is to really think. And this is also what we're seeing consistently now in the market. We have a. The CPA.com has a AI working group set up, an AI working group with practitioners and firms and vendors. And also in this group, we see this particular pattern emerge a lot, which is really your system of record is something you want to buy. You want to buy it smart, you want to buy it from people who are giving you powerful AI tools built on top of a really solid foundation. And then the little things which are idiosyncratic to your particular practice, those things are ones where I would think about, can AI help me make them better, faster, smarter?
B
Well, Pascal and Mark, we got a lot of questions coming in here which we're not going to be able to unpack today fully, but just a couple of comments on them. One, is AI going to replace tax software? Pascal was just talking about systems of record. It's going to evolve. It. There's going to be solution providers. All the solution providers are going to be leveraging AI for the tax engine, for the automation of the inputs. So yes, it's going to be, it's not going to eliminate it, it's just going to evolve it dramatically. And as I said, I think next tax season, you saw this tax season was about tax research. Next tax season will be about pretty much all areas of the tax practice. But this slide is kind of helpful. It kind of addresses a little bit of that question of like, where is the software categories going?
A
Yeah, absolutely. So, as we mentioned, and Eric, thanks for clarifying this, you have the SaaS solutions, the general ledgers, your tax research, et cetera. They will not go away. And you don't want them to go away. You actually want to use these things because they're really Complex, and they need to be, as you know, absolutely accurate. So you definitely don't want to bytecode this. On the flip side of that, as I mentioned earlier, the little scripts, little things which can make your life easier, the one thing I would encourage you all to try out a little bit more now is we're seeing these agentic desktop tools emerge. Claude has a product called Cowork. This is inside of their desktop app. You can give these desktop tools access to files and you can essentially program, as in directing, not programming, directing an agent to do things with those files. As I mentioned earlier, I had this recent use case. I had literally like 20 plus Excel spreadsheets. They all needed converting into a specific format. Instead of me doing this, I threw this all into Cowork, described to Corelor what I want to do, and then Cowork dispatched an agent, looked at the files and converted all of these files. So that is a pretty revolutionary step because it saves you a lot of time. I would encourage you to try these things out and play with them because again, particularly when it comes to AI, because it's such a fast moving and evolving area, it's really important that we create lived experience, that we actually build muscle memory in these things. It's not enough just to conceptually understand them.
B
Thanks, Pasco. Some questions that are coming in here, I'm going to throw one at you and you can just give me a broad answer. But where's the safeguards? Like, you know, what, what's going on with, you know, we've got all kinds of, you know, you got section 72 17216, where's the data? How are you managing data? What are the controls around that? Lots of questions, people saying how you, you know, where are you throwing all of this personal information? What are you putting into this?
D
Talk to your client about that. Like I said, the client's going out, they're probably in some free version and throwing their tax return and all their personal data up there without it. So if I'm the firm, the first thing I'm going to say to them was, congratulations, you just gave all your personal information in a deep, dark web, right, versus what the firms have. The firms are already in an isolated scenario, right? So whether they're in cloud and using the tax engines that are out there already now attaching on the enterprise versions of these AI solutions is going to be absolutely critical.
B
I hope regulation, but there will be regulation on, you know, you already have laws in place and there's going to be evolving regulation.
D
You know, you're gonna, we're gonna get the update in a little bit on AI and controls over that with Koso again where I say that we as a profession have a critical piece of that, but all of our members today, you better have that locked on. You better be in an enterprise version, not a free version. You know, I've seen some people say, why do tax research on Google and Gemini is good enough and they have no guardrails of understanding if that information is actually really up to date or if it's, you know, been preceded by other things that have come out. So it is important, I think, to be in a closed environment that you can utilize things and make sure that that data is safer.
B
Right. And so we're going to a summary slide here for Pascal and Mark. One of the tools that we have are the steps that a firm should go through in selecting the AI vendor. One of the most important things for firms to do right now is the due diligence and the proper selection of the tools that they're using. And you should, you know, we, some one question came in. Are you explaining that to the broader marketplace? I mean, that's something we can do and it's something that the firms can do with us.
C
Yep.
B
So Pascal,
A
all I have to say is truly you have to learn. There's this notion which is called the jagged frontier or the jagged edge of AI, which comes down to a gentleman called Ethan Mollick over at Wharton Business School. And he points out that the problem with AI is that it's not binary. It's not like it can do this thing or it can't do this thing. It's this weird beast which is ever evolving where it can do certain things really, really well and then it fails flat on other sides. So the only way you can actually get a sense for that is by experimenting. So my recommendation, three things. Try out these agentic tools. There are where the future will go. You will very clearly see much more of this agentic stuff. Adopt your package solutions strategically. This goes exactly to the point Mark and Eric you were making earlier. And please don't over invest into the custom build site. As appealing as it might be, if you're a little nerdy like I am, it typically is a dead end because it will just take you too much, much time. You will get a non standard compliance solution and it's often just the wrong turn to take on this road.
B
Well, Pascal, thanks for your partnership on this. One final question that I'm going to share that came in was what should you know, small firm, I'm a small firm. What should I be doing to kind of get on this in a cost feasible way? This is the great equalizer. There is solution providers that are putting in place the tools for the small firms. With that said, even a small firm can do some of the things that Pascal described today. This is an incredible opportunity for firms of all sizes.
D
Yeah. I mean the AI tool by itself individually is not all that expensive, whether you're talking about Claude or Copilot.
B
Copilot, a controlled enterprise version or if you're using Blue J for tax research and things like that.
C
Yeah.
B
All right, Pascal, we'll have you back in open forum.
A
Thank you so much.
B
So I bring up Danielle Huff. Great to have you out of our DC Office and looking forward to today's update and interesting to hear a little bit about the Direct File filing season from last year.
G
Thank you. So good afternoon everyone. We're going to start first with IRS operations. Tigta, or Treasury Inspector General for Tax Administration released a report on IRS's spending of those Inflation Reduction act funds. And of course, those funds have been whittled down over numerous bills, down to $26 billion. About 61% of that has been spent. So there's about 10 billion remaining. In 2025, the IRS spent about 6.7 of that. You can see the breakout by category on the slide. There's Next we'll go into the filing season stats. So far, IRS has received and processed slightly fewer returns than in the previous filing season. E filed returns slightly up. IRS website usage is still significantly up by over 60%. The average refund is still about the same as it was the last few times at about $350 more than the previous filing season. And then direct deposit refunds is up quite significant, significantly at about 7.3%, which is really illustrative of IRS's push for electronic refunds. Ticto also released a report on Direct file during the 2025 filing season. So of course we all know that Direct File has been suspended, but this report told us how widely used Direct File was. So Ticto reported there were about 32 million eligible taxpayers to use direct file in 2025. And of that there were only 751,000 registrants and only 311,000 users actually submitted a tax return using Direct File. On the next slide, we're going to cover tax relief for state declared disasters. So if you recall that HR517 was enacted last year that allowed states to request disaster relief. The governor must submit a written request to the Secretary of the treasury asking for relief due to a state declared disaster. And then the Treasurer, the Secretary has complete discretion to either grant that relief or not grant that relief. So there's only been two states so far who have received that relief. That's Louisiana and Tennessee. Last week, late last week, Tennessee received that relief for certain counties that were impacted by winter Storm fern that was back in January. And so those impacted will now have until May 22 to file tax returns, pay taxes for any tax deadlines that fell between January 22 and May 22. And then also we received some feedback asking whether states need to be proactively asking for disaster relief rather than waiting for IRS to simply provide it. And so we provided our state societies with a one page summary of this process and a template letter. And so this will give a little bit more explanation of that process and potential outcomes. On the next slide we have links to these resources. They're available for our tax section members. And also there's a link we have a webcast coming up on April 20th at 1pm where we'll be discussing disaster planning. Then on the next slide we wanted to talk about the Dirty IRS's Dirty Dozen tax scams. Just bring up a couple issues that are really relevant to our to our members. So IRS added back self employment on the next slide they added self employment tax credits back in due to misinformation about eligibility for those credits. They were actually Covid era credits but there's been a lot of commercials about eligibility and so just know that the IRS will be looking at the that will be closely scrutinizing these credits. There's also a reminder that scammers are out there. They're using spear phishing malware campaigns, trying to access your clients data, trying to access your systems fake charities. Just a quick reminder, use that tax exempt organization search tool to verify and then non cash charitable contributions. We just wanted to raise this because we've been made aware of a recent strategy where a team will promote an LLC or LLC or partnership interests that own a some sort of do good technology and some sort of intangible property. And what they'll do is they'll find investors, they'll sell blocks of the entity to them and then at the end of the year they're going to ask do you want to sell the asset, do you want to hold it or do you want to donate it? And the whole point of this thing is to donate that property. That property ends up being valued at about 5 times 10 times what the entity actually paid for it. And then taxpayers receive a charitable contribution deduction that is in excess of their investment. So the IRS has not called this strategy a scam or reportable transaction, but we're raising it because it is very the technique is very similar to what was used for syndicated conservation easements. For more info about what to look out for, we've also provided an article discussing too good to be true tax strategies. On the next slide, we're going to talk about penalty relief for farmers and fishermen. So generally, farmers and fishermen can avoid making estimated tax payments by filing and paying in full by March 1st. This year it's March 2nd. So on January 27th, IRS made some changes to Form 8995 and so tax which is used to calculate the qualified business income deduction. So software providers had to go in, take that form offline, make some changes and that affected the ability to timely file. And so AICPA submitted comments asking that the these the farmers and fishermen receive an extension from March 2 to April 15. On April 1, just last week we received confirmation that IRS provided that exact relief. The relief is automatic for those who file and pay by that April 15, 2026 extension deadline. Finally, we'll go on to our round Robin. On April 2, IRS issued another news release on those CP53E notices. And really it's just stating that for taxpayers that don't have access to their individual online accounts, you actually can call up and ask that and customer service, IRS customer service will allow you to submit a waiver if you are eligible for one of those waivers for paper checks. Our annual survey on IRS services is coming up after the 4:15 deadline. We ask that you please complete that survey. It really helps us understand IRS filing season performance practitioner pain points and really helps guide our advocacy efforts. And then last, I would just want to touch that. We also have our Washington tax brief coming up April 21st at 1pm where we'll dig a bit deeper into some tax issues. And we'll also be joined by IRS representatives who want to who talking about TaxPro data breaches and security. And that wraps up our technical tax session.
D
Daniel, thanks for that. And just to highlight for everyone, that AICPA IRS survey that is so critically important to us, we actually use that information. Daniel and his team, they're going to be talking to the irs, talking to treasury and we use that data just to keep them in tune to what's happening out there in the workforce around IRS related services. So once you come out of busy Season, please do take advantage of that. So, Daniel, thank you again. Now we're going to move on to our next topic, internal control in a Gen AI world. And I am excited to hear from Paul Perry and Lexi Weber. Lexi Weber is with us here at aicpa. Paul is one of our never tired volunteers and will do whatever we ask them to for the benefit of the profession. And we appreciate that. And I had mentioned already earlier today when I talked about Rise 2040 and this idea of trust, this falls right into that category that we need to as a profession, embrace coso, embrace internal controls, get it integrated, help business lines on the CFO business community, and then for the firms to come in and provide some level of assurance and advisory to their clients over the same thing. So I'll turn it over to Lexi and Paul.
F
Yeah. Thank you, Mark. And we're excited to have Paul back with us. He is no stranger to the town hall audience, so we'll get right into it just for sake of time. But Paul, you know, I think we're excited to talk through the new resources that COSO has provided the committee of sponsoring organizations of the Treadway Commission. But before we get into it too much, I was hoping you could give us a little bit of a refresher on what KOSO is and why it was originally established, especially for some of us who don't sit in the assurance space every day and may have not have. May not have reached KOSO in a while since our cpa.
D
Yeah.
C
So what you're saying is for those of you that don't have it as a poster in your office like I do, kind of, kind of give an update. Right. So back in 85, these organizations came together and it was really to address what that fraudulent financial reporting was that was happening in the US and when you do think about coso, you do think more assurance, but it goes beyond that. And as you'll see on the next couple slides, you really can apply it across the board for anything that a CPA does.
B
Right.
C
And I think that that is the biggest piece about coso and do I wish I had spent a lot more time in college, you know, writing a paper on this, probably. But once I got into the profession, and now that I'm focused on risk all the time, what it provides for everybody, not just assurance, but if you think about a tax strategy, if you think about trying to build an environment within an organization, think about the common language that is used to understand the uncertainties and understand the risks. And I think that that's the biggest piece, when you look at internal control, integrated framework, icif, and I've had the chance to talk to some people who were a part of coming up with some of the conversation that's. That's in these frameworks and guides. It really was, let's create a common language we can all use, no matter what it is, to get through a decision that needs to be made, whether it's controls that we need to verify that fraud's not occurring or its controls around AI, which we'll talk about in a minute. You know, we always say trust, but verify. And I was on a webinar earlier today and somebody used the term never trust, always verify. Right. And that's part of what this language does for us is it keeps the human in the loop, in the lead, however you want to say it. Around internal controls, we can put in technology, we can put in software, but it's still the human that needs to be apart. Verify, understand what's happening. And that goes back to that trusted advisor that Mark is always talking about at everything that I've ever heard him speak on. And Pascal was even talking about it too, as it relates to AI. And the other thing it does for an organization is if you sit down and you say, okay, what are our risks? What are our uncertainties? It brings everybody to the table. Right. You don't want to do risk analysis, risk understanding in a silo, in a vacuum. You want to do it with everybody involved. And I think that that's the biggest piece of. On why it matters to everybody within this profession.
F
Yeah, absolutely. I couldn't agree more. And I think we've seen that. Right. In different conversations that you and I have had about the concern and the AI anxiety of what those risks can bring. And as we move to the next slide, you know, one connection I really wanted to make was that we did see that in our epicenter webcast that we were on last week about discussing risks around automation and AI. There was a lot of concern of, you know, what, what am I supposed to select kind of to what Pascal was saying earlier. You know, what systems do I use? How do I know it's safe? What information can I put in it? So I think this is a really timely report to discuss. How can we use this to really guide us through applying the COSO internal control framework into oversight over gen AI? And how can we really address those new unique, take risks and implement these artifacts into our everyday conversations?
C
Yeah, and you could, you could spend, you know, 30 minutes on this. So I'll High level. It, it's not a new framework, right? It's taking ICIF and it's applying its use around how do you harness Gen AI responsibly? And I think what Pascal, if I was reading between the lines, and I agree with all that he said, one of the things is there's FOMO in, in society and in business. Everybody else is doing this, everybody else is using this model. Why are we not? And so they just go forth and they jump in and they start using it and they haven't thought about the governance and they hadn't thought about the controls. And so what, what Koso's framework, ICIF framework and this guide on internal control over Gen AI does is it steps back and it says, okay, the one unique thing about, you know, you could think of COSO as kind of a pros cons list, right? This is the I want to do something. But these are the pros, these are the cons. What COSO brings into the play is controls. Somebody asked that question at the very end of Pascal's session, you know, how do we do this correctly? How do we do this responsibly? And it's bringing in the controls. And the one unique thing that Koso does provide is kind of thinking about risks and thinking about uncertainties. And now we have a whole new set of risks that five years ago we were not even talking about hallucinations. Why would you have ever considered hiring a person who lies on their resume, who always makes hallucinations in their business? Right? We would have said that's not part of our culture, our understanding. And so now we have AI and no offense, it does that sometimes. And so we have to put the right controls in place. And so if you're going to pick up this document, go kind of middle of the document and there's a good six step implementation cycle, starts with governance, ends with monitoring, follows the COSO cube correctly, perfectly. And I think that cube can be established everywhere. Set the tone, what are my risks, what are my outcomes? What controls do I have in place to make sure I have the right outcomes? And then how do I communicate that and how do I monitor that, how do I adapt that? So a really good guide that was created for those that don't know how to implement Gen AI within their organizations. It's a good place to start. Good read.
F
Yeah, absolutely. And I think too just getting that starting point of we know where all these things exist, but I think it's nice to have it laid out in a platform that you can really Structure conversation around because I think that's something that a lot of people struggle with and kind of the overwhelm of what AI can bring to the table. So as we move to the next slide, I also wanted to touch point on the new guidance around corporate oversight, corporate governance and board oversight and how everyone can really apply this into their organizations, regardless of the size or structure. And I think that's one of the really applicable formats that we can do it. And you know, we always talk about tone at the top, especially going back to our accounting days. Right. That's more important now than ever, I think, as we implement AI and kind of navigate through this cycle.
C
Yeah. So this document, again, a guide, a publication more like guidelines than rules. We're not trying to say thou shalt do X. It's if you want to do good corporate governance, if you want to have good oversight of your organization from a board perspective. And again, it's not the size of the organization, it's anybody can use it. But if you have something that is like a board that is governing your organization, how do they set the tone? And I think one of the biggest pieces that I pull out of this document is culture and purpose are a control. And people don't usually think of them in those terms. And I always say go back if you have an issue, if you have a problem, if you have a significant deficiency, if you, if something was missed in attack strategy, go back to the culture of the organization. Go back to the culture of whoever was establishing governance and say, did they allow that? Did they suggest that? Were they suggesting something else? Right. Do the right thing always is a great motto for folks. But I think if you create culture, treat it like a control, treat it like something that is a good idea, then it's really, it should help your organization from a governance perspective. There's 12 guidelines, there's guiding principles over four areas. It also is a really good read. It's a short read. It's only about 40 or 45 pages. But if you are need to start somewhere with governance, this is a great place to go and start.
F
Yeah. And I think it was nice when I was looking through the report and reading it. There's four main themes that they really call out that we have on the slide. And I think it's nice to kind of bucket. Okay, what pieces do we need to apply for each bucket that we would apply it to? So, you know, I think this is a great conversation to have. And Paul, of course we always appreciate you on town hall and willing, being willing to jump in, discuss your areas that your ex you have expertise in. So as we move to the next slide, one thing just to continue the conversation for everyone on the call, we wanted to provide some additional resources, especially if you're, you know, just starting to reintroduce yourself with the KOSO framework and these new guides to help implement into your strategy as you're looking into general generative AI and other emerging technologies. And we also put some AICPA resources as well on ethics around AI and just some glossary of terms as you're getting familiar. So thank you, Paul. We really appreciate you.
C
Thanks, Lexi. Thanks, Eric. Thank you.
B
Yeah, thanks Paul. And Lexi, you're staying with me. That was a great segment. Just a perfect follow up for our discussion in AI because there was a lot of questions on how, how we're going to kind of lead with controls and risk management related to gen AI. So thank you, Lexi. So first off, I know you're going to engage. Well, I'll be there. And we're actually going to kick off engage with the AICPA town hall. But then following us on that first day, Monday, June 8th, you're going to have a discussion between Ryan Reynolds and the former Costco cfo. This is something that they, they've been collaborating on, so it should be very dynamic and insightful. And then we've got Ian Bremmer who will be talking about the overall global affairs as well as markets. And that's just a few of the presenters. There's many, many more and we'll be highlighting more of them in future town halls.
F
Yeah, I'm excited for it. I think it's a great lineup and a tradition like no other in light of the Masters week for the AICPA. So it'll be exciting for our 10th year. You know, we also just wanted to be cognizant as well and provide some resources to the audience on some concerns around cybersecurity as well as business operations and how those are currently being impacted by the conflicts in the Middle East. So there's some resources here for our audience just to help navigate that and just help you kind of understand what is out there to help with those risks to address that.
B
Thanks, Lexi. So just a quick click through here. I know it was about a week left before the first filing, you know, tax filing deadline. In here. You still have the opportunity to leverage this tax research tool with a seven day free trial. So let's start, let's move to the open form. We just got a couple minutes here, but do have A good question, Rachel, for you. And this is related to tariffs. We didn't address this today, but what is the status with the rollout of that tariff refund platform, which has got this acronym cape of all things?
E
Yeah, so the status of that is the courts said that the administration has to roll out their system for refunds by April 20th and they have to roll out what they're planning to do by then, but they want a status update by this next Tuesday, April 14th. So we'll know more on Tuesday, but it is in progress and it is occurring this month, so we'll hear more about that. And then also just on the issue of tariffs, there will be a hearing tomorrow on the Section 122 tariffs, so we're following that.
B
Thank you. All right, so we're in lightning round here, Pascal, but a lot of questions came in just on these enterprise versions and the safety of the data. So I know you know a lot about this, but just maybe comment on that.
A
Yeah, rule of thumb is if it's free, don't put your data into it. Very simple. The moment you start paying for it, you should do yourself a favor. So, so typically the moment you start paying for, let's say chatgpt, Claude, whatever it is, you are more secure, your data is more secure. But do yourself a favor, read the terms and conditions pro Trip because they can be very long. You can actually put them into AI itself and ask AI about the stipulations in there and save yourself a lot of time. But yeah, please do pay for.
B
And then Pascal, we do have that toolkit as well that they can leverage just on the steps to do evaluation. So thank you, Paul. Thank you, Lexi, Rachel, Danielle, Pascal, I've just got a couple of closing slides here, but greatly appreciate you being on today's town hall. So next town hall, we've got a Great lineup for April 23, DC update, technical updates and we're also going to be doing a deep dive related to audit transformation. We are launching a new survey today. I mean the old survey at the end of the town hall used to I think have like 30 questions. We still have the NASBA CPE related questions, but we've got three or four questions we would love you to answer. Your feedback is very helpful. So please take two minutes and do this post town hall evaluation for us and we do have the additional resources here@cpa.com townhall it was great being with you today. Have a good end to tax season and for all of you, greatly appreciate you being with us. Thank you very much.
E
Thank you for your participation. You can also subscribe to the AICPA Town hall series on your favorite podcast platform, as well as watch archives on
A
YouTube and find resources@cpa.com Townhall Tune in
E
for live broadcasts Thursdays at 3pm Eastern Time.
D
This podcast is designed to provide illustrative information with respect to the subject matter covered and does not represent an official opinion or position of the AICPA or AICPA.org it is provided with the understanding that The AICPA and AICPA.org are not engaged in offering legal, accounting, or other professional service. If such advice or expert assistance is required, the services of a competent professional person should be sought. The AICPA and AICPA.org make no representations, warranties, or guarantees as to, and assume no responsibility for the content or application of the material contained herein, and especially disclaim all liability for any damages arising out of the use of, reference to, or reliance on such material.
Date: April 9, 2026
Hosts & Guests:
This episode provides a comprehensive update on the state of the accounting profession, key economic indicators, rapidly evolving AI technology and its impact on the industry, technical developments from the IRS, and crucial updates and guidance on internal controls in the era of Generative AI (GenAI). The discussion spotlights both present challenges and forward-looking strategies for the profession to remain trusted advisors, with a focus on leveraging technology while maintaining ethical and regulatory standards.
[02:10 – 09:26]
[20:47 – 36:43]
[32:30 – 36:43]
[13:44 – 44:44]
Washington Update:
IRS Operations:
Tax Scams & Charitable Contribution Risks:
Penalty Relief for Farmers & Fishermen:
[46:09 – 55:24]
“We are still the trusted profession in business—making sure that we maintain that level of trust… That’s what we were built on.”
– Mark Koziel [05:26]
“Can we build it? Should we build it?”
– Pascal Fannette [27:17]
“If it’s free, don’t put your data into it. The moment you start paying for it… your data is more secure. Read the terms and conditions—pro tip: you can put them into AI and ask about stipulations.”
– Pascal Fannette [58:39]
“You want to do risk analysis… with everybody involved. That’s the biggest piece of… why it matters to everybody within this profession.”
– Paul Perry [48:28]