The 24th Annual Financial Reporting Conference, hosted by Baruch College, New York, N.Y., on May 7, 2026, provided a forum for policymakers from the SEC and FASB to sit down with financial statement preparers, accountants, and business leaders to talk about current topics in business, accounting standards, and enforcement.
The opening session featured remarks from FASB Chair Richard Jones and SEC Chief Accountant, Office of the Chief Accountant, Kurt Hohl. They discussed their respective institutions’ strategic priorities, especially in light of the new leadership at the SEC. Hohl mentioned the need for an updated inspection process and meaningful reports. He also discussed the opportunity to streamline the 10K in the context of the SEC’s proposal to move to semiannual reporting, remarking that if investors have too much information they might not take it all in. Jones noted the value of standards setters in taking a holistic look at the most critical information being reported by filers.
The next panel, “Current Developments at the SEC,” featured SEC representatives explaining their work with registrants and expressing their concerns about the issues they see in enforcement cases. Moderated by Sheri York, deputy chief accountant, Office of the Chief Accountant, SEC, the panel featured Heather Rosenberger, chief accountant, Division of Corporation Finance, SEC, and Ryan Wolfe, chief accountant, Division of Enforcement, SEC. The morning concluded with a panel discussing hot topics at FASB. Moderated by John Vanosdall, partner, PwC, the panelists included Jackson Day, technical director, FASB; Mark LaMonte, partner, Williams Marston; and Lara Long, managing director, Riveron.
The luncheon speaker, Howard Schilit, author of Financial Shenanigans, sat down with Conference Chair Marc Siegel and discussed the various ways that businesses try to manipulate their financial results and the public scandals that have followed (e.g., Waste Management) when such practices have eventually come to light.
The first panel of the afternoon focused on audit committees. KPMG Partner Kimber Bascom moderated the panel, which included Mark Besca, board member at Markel Group, Las Vegas Sands and Clarus Corporation; Robert Herz, former FASB chair; and Peggy Smyth, independent board member. The conference concluded with a panel that covered how investors and analysts use financial reporting information. It was moderated by Neri Bukspan, co-founder and executive advisor, Metrix Advisory, and featured Jeff Brickman, FASB senior investor liaison, and Todd Castagno, managing director, Morgan Stanley and EITF member.
The afternoon’s agenda also saw a panel discussion on artificial intelligence, moderated by EY Partner and Professional Practice Director of ESG Matters Jason Bond, and featuring Michael Cerami, EVP, Strategic Alliances, Business Development and Communications, CPA.com; Denise Lucas, IBM vice president and assistant controller; and Jonathan Rhine, PwC Assurance Innovation & Technology. The following article is an abridged presentation of this panel discussion, edited for clarity and length. The views expressed are the presenters’ own and do not necessarily reflect the views of their employers or affiliated institutions.
Artificial Intelligence
A Panel Discussion
(Left) Panelists Jason Bond, Michael Cerami, Denise Lucas, and Johnathan Rhine. (Top) Jason Bond and Michael Cerami. (Bottom) Denise Lucas and Johnathan Rhine. The panelists discussed their experiences with and predictions for AI in finance and accounting.
Jason Bond: You can’t go to an accounting conference these days and not have a panel on AI, but we really wanted to try to make this panel a little bit different. We’re going to try to focus on what it can do for the profession and what people and firms can do to enable that transformation in their organization rather than presenting a primer. We’ll share some real examples, not just hypothetical learnings. Let’s start with strategy, because most organizations jump straight to the tools without first deciding what they actually want to change. Jonathan, your vantage point allows you to see AI across the industry. When organizations say that they’re adopting AI, what separates true transformation from disconnected initiatives?
Jonathan Rhine: I’m going to draw a distinction between the tremendous amount of AI activity we’re seeing and what I’m going to refer to as transformation. By now, I would argue that every employee has a chatbot, and that’s table stakes. At the same time, you’re probably not seeing a change to the underlying business model. If you’re following the same business model you’ve optimized, you haven’t necessarily transformed. Optimization asks, “How can I do whatever I’m doing better, faster?” Transformation asks, “What activity should I actually be performing?”
Peter Drucker had a theory that every business is founded under certain core assumptions. Who are their customers? Why do their customers come to them? What’s their competitive advantage? Who’s their market? Drucker said that companies don’t generally fail because of poor execution. They fail because the underlying assumptions changed and either nobody noticed or nobody did anything about it. AI is challenging assumptions at breakneck speed. Inherent in so many business models is this assumption of limited capacity, of time and resources and individual work, a limitation around information. I think organizations that are just optimizing old business models could be at real risk.
At the same time, there’s a tremendous opportunity here to ask the question, “What’s the new task?” What’s the new value that we want to add?
What I find most effective is answering a single question: what problem are we actually trying to solve? You put all of that into perspective, and now you have transformation activities that are focused on that specific objective.
What problem are we actually trying to solve? You put all of that into perspective, and now you have transformation activities that are focused on that specific objective.
Now, some people are probably thinking to themselves, “That’s great, for someone else in the organization, that’s not for me.” I’m going to challenge and push back on two points. The first is that this is for everyone. Everyone in every organization has to think bigger. The second element is that finance and accounting sit at the intersection of every business transaction. The finance function is perhaps one of the most comprehensive sources of information within an organization that might be needed to drive the AI transformation.
I’ll close with just two things to think about. First, what assumptions are inherent in your organization that AI is about to change? And second, what could your organization do now that it wasn’t able to do before?
Strategic Questions
Jason Bond: This really sets the stage for thinking about strategy. Denise, hearing that framing as a company, how do you think about AI strategy?
Denise Lucas: At IBM, we think about our enterprise AI strategy from both the top-down and bottom-up perspective. At the top, you need clear priorities, strategic investment decisions, and clear governance. We also have top-down corporate initiatives where we pool corporate resources, have cross-functional collaboration, and reimagine workflows to focus on the transformation. Where do the best ideas sometimes come from? They come from the people who are doing the work. They know where the friction is. They know where manual processes may not be working.
We do want every IBM staff member at every level of the organization to be upskilled and able to innovate and submit ideas that can transform the organization. We have delivery engines and both those corporate initiatives that are kind of large cross-functional work stream transformations, and playbooks that give practical insights to domain-led teams, and that drives transformation at both ends of the spectrum. We also have what we call enabling motions, and those could be anything from partnerships to culture, as well as governance. We call governance an enabling motion, not to slow the business down, but to enable speed and transformation in a trusted, transparent way.
Jason Bond: You touched upon governance, anything else you want to add about how companies think about governance as it relates to AI transformation strategy?
Denise Lucas: Because of my accounting background, part of my role is to focus on domain-led transformation in particular use cases that could impact financial reporting, directly or indirectly. I think we’re inherently trained to identify risks and then identify controls to mitigate those risks. Governance is embedded in the profession, as governance is key to transformation. It enables people to act with confidence.
I also think that governance should be scalable. It should be a risk-based methodology. A personal productivity use case of generating a first draft really shouldn’t have the same level of governance as an end-to-end workflow transformation that could impact financial reporting. I think governance needs to be clear and scalable throughout the organization.
Jason Bond: Michael, some in the audience may be thinking, “What Jonathan and Denise described works for big organizations that have significant AI budgets, but how should a smaller company be thinking about AI strategy and transformation?”
Michael Cerami: I completely agree with Jonathan and Denise. I think they make great points about the distinction between real transformation and process optimization. But I’m here to represent the voice of the smaller company or firm. Many of those small businesses are looking to firms for help figuring out how to integrate AI into their strategies and what they do.
We’ve created an AI working group. We meet monthly, and it’s a combination of chief AI officers from some of the firms here in this room, from midsize firms, also from smaller firm practitioners and CEOs, and then we sprinkle in some of the providers themselves, their CEOs, and founders, and just talk about how all this is coming together. I joked with Jason that we meet monthly, and I can’t ever think of a topic that’s changing so fast, so quickly.
I think the small firm practitioner sees AI as having the potential to be the great equalizer.
“How do I compete?” My biggest challenge is building capacity within the firm, doing more with less. I think the best way for those folks to see how this can be a game changer for them is to just get on the field and start testing some of this stuff.
I think a lot of that is beginning to happen. They’re surrounded by AI tools today that can make a pretty big impact. And I think that builds more confidence that leads to them taking a step back and thinking more broadly about their business model. It could be everything from pricing to process workflows.
Culture and Change
Jason Bond: As I was preparing for this, I actually saw a couple of studies that there’s cultural hesitancy to use AI among a lot of employees. One survey identified a significant disconnect between executives and employees, with 54% of workers bypassing company AI tools, and 33% not using them at all. Denise, what have you seen that has worked well when it comes to training, incentives, or norms?
Denise Lucas: I would say that it is as much of an organizational change management effort as it is a technology effort. The technology may be new and continuously evolving, but to get people to adopt, you need trust, leadership behaviors, incentives, and to convince people that it’s going to make the way they work better. We’ve been focusing on the cultural aspect—that enabling motion that’s cross-cutting throughout the company.
In terms of norms, I think it goes back to what Jonathan said originally. The norm that we’re trying to create is don’t use AI for AI’s sake and just check a box. Think about the problem you’re trying to solve. What is the business objective? Then what is the workflow that you’re looking at and how can you make that better? At IBM we look at it in an eliminate, simplify, and automate way. What can we eliminate if it’s not adding value, what can we simplify if we over complicated it, and then how does the technology help solve the business problem? I think that’s helped in terms of getting our teams to think in a way that’s not just “We have to use AI.” Some of the best practices that we’ve seen and things that have resonated with our teams are created in communities of interest. To bring together that bottom-up layer, we have champions who we use to get out the word on specific topics. I also think enablement and education need to be consumable and simple, so recently we enhanced a landing page, our one-stop shop for our finance and operations functions which is open to the whole company as well. The idea was, “How can you get all the information you need to know about deploying AI and governing AI within a three-click mentality?” So that’s been super helpful as well.
Jason Bond: Michael, what have you seen that has worked to change day-to-day behavior, particularly with accountants?
Michael Cerami: I think you have to be very hands-on to drive change and behavior. I can’t emphasize that enough: be thinking through quick wins that you can begin experimenting with in a safe way. I think Denise said it wonderfully well: The tech is really not the challenge, it’s all of the change management that lives around the tech—the workflow, the process, just changing behavior. We’re going through something similar working with some firms through that transition and it is a very big endeavor and requires so much up-front buy-in. The technology is changing at breakneck speed, and to get human beings to change just as quickly requires some real work.
Jason Bond: Jonathan, do you have additional advice for improving the cultural change and how companies may think about return on investment of AI transformation?
Jonathan Rhine: I’m going to take that in two parts. We’ll hit first on the cultural and then the ROI. On the cultural side, I think Denise and Michael made some extremely well-articulated points. The step any organization needs to go from providing a product that is now available to the employees, to them actually using it is extraordinary. I’m a big believer in hands-on experience. I found myself with a tool that I should have been using because I created it, and I caught myself one day actually doing it the manual way. And I thought, “If I can forget to use the tool that I created for the organization, what are the people out there doing?”
It’s hard to change. One thing I find particularly helpful is hands-on workshops. You’re in the room. You’re building something that’s genuinely useful. I personally find it breaks the mental block, because you’ve seen it once and now you’re like, “Oh, I could do more with that.” Make people part of the solution. We’re not here just to give things. It’s a collaborative model.
Switching over to the ROI question, I’m going to go back to the original comments I made about transformation. There are many, many unknowns. So ROI, by definition, is going to be difficult. If you’ve clearly defined your transformation objective, then at least you know what you’re trying to measure. If you say I’m going to cut costs, then you’re going to try to measure costs before and after. If you say it’s about improving quality, maybe it’s hard to measure quality, but at least you know what you’re trying to measure. Make your strategy your scorecard. At least that takes you one step closer toward the measurement.
Jason Bond: Denise, how have you thought about ROI and maybe measuring it differently than you would in other types of projects?
Denise Lucas: I will repeat what you said, but I think if ROI doesn’t start with purpose, you risk measuring activity over impact. I think when this all started the focus was all cost savings and time savings. It’s still part of the scorecard. I do think there’s a shift towards opportunity and trying to measure that value. Accuracy, quality control, risk mitigation, and user client experience are all really important. I think a scorecard can have both quantitative measures—the cost and time savings, adoption rates, accuracy, avoided costs, financial return, and all of those numbers—and it can discuss the intangibles or those qualitative factors. Is it changing behavior the way you wanted it to? It doesn’t need to stay static. See what’s working and resonating, the data that’s available, and improve it over time.
If ROI doesn’t start with purpose, you risk measuring activity over impact.
The Data Underneath
Jason Bond: We’ve talked about strategy, we’ve talked about overcoming cultural hesitancy, but then when you actually get into AI implementation, one of the pitfalls is that organizations haven’t really focused on the underlying data. Jonathan, when we were preparing, you’d mentioned that some people think that they can just throw an AI tool on top of data and it will work magic. Why is that assumption dangerous in the accounting and financial reporting area?
Jonathan Rhine: I think a lot of people view AI like this magic tool. I’m just going to throw it at whatever my problem is, and it’s going to solve it. But I want to drive home: AI doesn’t make your data problems disappear. AI makes them impossible to ignore. I think that’s a really important concept. Whatever the reason, an inconsistency here, a gap there, we all understand what goes on in an organization. AI amplifies those at scale. If you’re using AI with that data, AI is going to magnify whatever is there throughout the decision-making process going forward.
I like to call it my litmus test. If a human was given all of that data and had an unlimited amount of time, but no prior knowledge, would they be able to infer from the data what it is that you’re asking AI to infer, to come to those conclusions, to reach those analyses? If a human with all of that extra time is going to find themselves confused by the way data is organized or perhaps is going to find missing information or inaccurate information, AI is going probably going to hit those same roadblocks. If a human can derive it from that information, AI is able to operate a lot faster and it’ll be able to extrapolate that information and pick out the pattern that we’re looking for.
The finance function, the accounting function, sits at the intersection of business transactions. That data is the data that many of us here in the room are involved with, and responsible for. An investment in data infrastructure and the finance function might serve the AI ambitions for the entire organization.
Jason Bond: Denise, can you share how you’ve approached data readiness with the AI tools that you were looking to implement?
Denise Lucas: When I think about data, I think it depends on the type of solution or type of tool that you’re looking to build or use. For the consumer-like LLMs that are foundation models and built on massive amounts of data, the onus really is on the end user. If you’re building a tailored model and leveraging trusted data sources, your approach may be different. I spend much more time with my chief data officer than before, and I’m learning that some of the data considerations of the past, like data fields and lineage, are now getting more complex. Looking at your sources of data in that context is helpful. I also think data classification and tools are important, especially when you’re using third-party tools. AI readiness isn’t just technical readiness. It’s data readiness. It’s user readiness. It’s process readiness. It’s control readiness. If all of those things aren’t aligned, then you can have a super impressive tool, but the business impact is going to be limited.
AI Agents
Jason Bond: There always seems to be buzz around different aspects of AI, and recently a lot of this has been the power of agents. As a reminder, an AI agent is a software system that can perceive information, but then it can make decisions and even take actions to achieve a goal. That may be across multiple steps and multiple tools, even without needing human instruction. Michael, how do you think AI agents will change the accounting and financial reporting profession?
Michael Cerami: What I’m seeing is a lot of the established platforms and tools out there building agents into their systems that firms and users can leverage. There’s a lot of progress and investment being made in closed agents. For example, anomaly detection, a kind of assurance agent that can flag stuff and bring it to the human’s attention in a much quicker way. Even financial intelligence, where you can ask specific variance analysis questions of the data in the platform is an example. With closed agents in particular, there’s been a lot of investment, a ton of options built into existing tools, or living outside of existing tools, that can speak to big ERP systems and create lots of efficiency.
Jason Bond: Jonathan, can you share how you’re seeing the power of agents come to life? I know we’ve talked about deterministic versus probabilistic agents and how they work together and how that’s really transforming the task that may be undertaken.
Jonathan Rhine: The key difference between simply using an LLM and the power of agents is that you’re asking AI to figure out how to accomplish a given task, and you’re giving it the resources to be able to do that. We got the ability for that LLM, that foundational model, to act with intelligent coordination, to take a problem and for it to use that same logic that it uses when it converses with you in order to be able to direct other resources when necessary.
A concern that I often hear, especially when working within our profession, is that we’re looking for deterministic outcomes. We want to be able to have a process that follows precise rules: specific input, specific output, and it runs the same way every time. In other cases, we want more probabilistic behavior, we’re looking for the softer skills that an LLM can provide. The power of that orchestrating agent is that it can look at a situation and decide what needs to happen and make sure that the right tool is used.
The impact that agents are going to have is incredible. I think we all know it. But understanding what’s going on under the hood and making sure we’re using the right tool for the right job is essential, because that’s when it’s going to drive the outcome and the quality of the answer.
Jason Bond: What kind of oversight are people performing over agents? How are we mitigating risks of hallucinations and making sure there is oversight of the output?
Jonathan Rhine: Last year at this panel, there was constant conversation about making sure that the human had to replicate every single step of the way. As our thinking and agents have matured, we’re thinking about when it is appropriate to reframe that and start thinking about what people refer to now as “on the loop.” Humans are still involved, and still taking responsibility, but if we’re using certain processes, are we in a situation where we can define the parameters and we can have a human step in when something falls outside the parameters? Michael gave the example of a closed agent. Do we have the possibility where routine reconciliations are hardwired? That’s not something that we need to address, the human could potentially be there just to step in when something in the closed process doesn’t reconcile.
We need to gain comfort. We can’t go blindly. At the end of the day, it’s our responsibility. And we need to think about how we retain our skills. As AI offers more and more in terms of insight and judgment, how do we maintain our own domain knowledge? How do we make sure that we can exercise our understanding and our skills when necessary and we don’t lose them or defer to AI?
Jason Bond: Denise, you talked about internal controls. What do you think about governance oversight controls over the AI agents you’ve seen?
Denise Lucas: I think about this a lot. As everyone alluded to, going from reviewing output to action and potentially autonomous action really does change the risk profile of the workflows. I think this also drives the need for a shift from policy-driven governance and manual governance to using technology to govern the technology. If we’re using end-to-end workflows that have access to critical systems or critical data, we need to have monitoring, logging observability, and the ability to intervene instantaneously if something goes off the rails. I think it’s shifting away from more manual-driven governance to something akin to cybersecurity and using technology to understand how the technology is acting and then put controls over the risks that are within the workflow.
At the end of the day, it’s our responsibility. And we need to think about how we retain our skills.
Jason Bond: If you’re like me, you’re probably seeing articles all the time that talk about how AI is going to change skills and duties, particularly for those white-collar jobs. In your opinion, two years from now, what will a high-performing accountant spend more time on, less time on, and what will they do for the first time?
Michael Cerami: More time spent on judgment calls, client advisory, and reviewing AI outputs. Especially in the middle and down market, it’s really interpretation and insights around what the data says versus the compilation of it. Particularly when you go to mid- and smaller-size firms, they’re just wired that their value in the relationship is the building of a lot of the content and working on mundane tasks. I think if they can be reviewers of that information, but a lot of the focus becomes placed on guiding that client, I think it’s a great outcome. It’s workflow, process, business model change. And so that’s what I think needs to happen, and will happen, over the next 18, 24, 36 months.
Denise Lucas: I would say subject matter expertise will become more important, not less. Building relationships and working cross-functionally to put initiatives into action will also be more important. In terms of less time, repetitive transactions pulling data reconciliations. I think high-performing accountants for the first time will be truly redesigning AI-enabled workflows.
Jonathan Rhine: I’m going to echo what people said, with my own spin. Definitely more time on client relationships and strategic counsel. More time developing the vision, the entrepreneurship, the dream. The AI is executing for us—we need to dream about what it is that we want it to do. Less time spent on data gathering, reconciliation, routine reports, answering questions that have clear rule-based answers. For the first time, I do believe that there’s going to be increased expectations. Do we now find ourselves with an expectation of more procedures, more to deliver? Whoever our client is, whoever our stakeholders are, we have that capability now.
Jason Bond: What does that do to the role of the future accountant?
Jonathan Rhine: The key message is that AI raises the floor, and humans raise the ceiling. What I mean by that, again, is the basics are getting automated. The bar keeps moving up. The ceiling, though, that’s the vision that we have. That’s our role to execute.
Where does AI fit in? We go ahead and we have the domain knowledge. We leverage it. We think of it as a team member. We know its strengths. We know its weaknesses. Ultimately, the scarce resource is humans. People are going to have AI. What they don’t have is someone who can exercise that kind of judgment, who can exercise those people skills, who can have relationships, and truly have the vision. The profession does not become any less relevant. On the contrary, I think we become more relevant, because we become more empowered to exercise the things that only we uniquely can—we matter more, not less.
Jason Bond: As we close out the panel, I’ll let each panelist say one point they want to emphasize that audience members should take away from our discussion.
Jonathan Rhine: The organizations and the professionals that are going to thrive in this era are not necessarily going to be the ones that have better technology. I think it’s going to be the individuals and the organizations that are willing to challenge their current assumptions, challenge how value gets created, do the hard work of reimagining their business, and never lose sight of what they as individuals and as people can bring to the table.
Denise Lucas: This is a really exciting time. People are so excited, and it’s invigorating to be around. I am an accounting lover at heart, so I do think that the accounting work that’s most valuable will become more important, but I think accountants will also become more important. They’ll be the trusted business advisors that people work hand in hand with versus just closing the books and recording the transactions.
Michael Cerami: I don’t think trust and ethics have ever been more important, and that’s what this profession represents. I go back to the adaptability question. Unfortunately, I work for a company, we love to do presentations, I hate doing presentations, so I started using Claude to build my presentations. It has saved me so much time, and you should begin to apply that to other parts of either your personal or professional life, and see how you can optimize a lot of things you do. I think that’s what really continues to get you excited and enthusiastic.
