Name: Finance Forward 2025 - AI Public Company Playbook
Uploaded: 2026-01-22T03:46:04.902Z
Duration: 30 min 8 s
Description: Finance Forward 2025 - AI Public Company Playbook

Transcript for "Finance Forward 2025 - AI Public Company Playbook": We're ready to rock, your minds and talk about AI and public companies and, how these three legends operate, under some pretty intense scrutiny. So I'm so excited today. I'm Christine from Excel, to be joined here with, Sekit, who's the CAO at Fastly. He was previously at Qualys, Marin Software, and Accurai. So he's had a lot of really great experience. We then have Micah, who is the AI principal at PwC. He has a ton of great experience from Udemy, Meta, and PwC, working with a number of really fantastic companies. And last but not least, we have Patrick who is the CFO at Lima One Capital. He has extensive experience leading finance and accounting functions in public markets. So round of applause for these three incredible leaders. Great. Well, maybe to get started, let's talk a little bit just about the moment in time that we're in. I think every session today is going to talk a little bit about AI, and that's the moment in time that we're living in. And it's just really exciting to see all the innovations. But I think public companies face a lot of increased scrutiny and regulations and compliance requirements that other companies may not. Patrick, do you wanna share just a little bit about the the biggest opportunities that you see, that that companies can leverage AI in in this time? Yeah. I mean, it's right. Being within a public company, right, Sarbanes is absolutely massive for us. Right? And we talked through comes really about the guardrails of controls or, you know, impact of controls, and and that's present in our mind every single day in how we operate. Right? So we have probably excess compliance costs because of that. I'm really looking at AI as an opportunity to really reduce that cost load, and get a little bit more efficient and and more timely in how we operate while still doing it within that controlled environment, right, using those those guardrails to protect us. We are in process of implementing Campfire, and it was kind of in that pretense that we selected Campfire was based on having the appropriate control environment workflow approvals that we need to operate, and to do so within, you know, a a better tech platform than where we were currently. Currently. Makes sense. What about regulations, though, and some of the compliance requirements? What do you think the biggest questions or concerns you kind of have to deal with on a daily basis when you talk about implementing some of this AI technology? So we get it from a couple places. We're also so let me once say, business purpose mortgage lender. So we've also got mortgage compliance sitting on top of any SEC compliance that we have sitting there. Again, it's controls first for us of whatever regulatory oversight needs I have. We're looking at, okay, where are the controls to protect us within that framework? And kinda getting back to, great, how do I operate within this environment? Again, I know I'm gonna have additional overhead. I know I'm gonna have additional hurdles that have to get over, but how do I do it in the most efficient way possible? We a good example is we implemented RAMP earlier this year. And while, yes, we used it in the thought primarily of how do I get better accounting information through and financial information through to help manage the the spend of the business, We also quickly realized that we had just a lot of frivolous spend and lack of control around spend from our prior, card platform. And some of that was actually we had some compliance issues in that that ramp was easily able to give us insight into and give myself and my team insight into where we could kinda nip that immediately, and make sure that we didn't run risk of potential regulatory issues. Makes sense. Micah, you've worked with a number of different companies of all different sizes. What do you think the biggest misconceptions are when folks talk about leveraging AI for their finance organizations? So I think there are probably three I would call out. So first, companies believe that it's not possible to move quickly or realize benefits because of SOX. And a second is, we can't adopt because of our kind of the quality of our data or the data that we have available. And third is kind of this idea that we can't adopt because we have the wrong technology. So if I just take those in order, think for the first one, it's interesting because when you think about SOX adoption, there are a number of ways that companies are able to do this effectively. And when you look at leading companies, they are already in the process of scaling, like the use of agents in finance, for example. Yep. And there's lots of ways that you can do it, but, you know, that frequently then people can get to that second barrier of, hey. Like, we don't have the right data or we don't have or we don't have strong data governance across the board. Where do like, what can we even do? Yep. The the problem with there is that every single company has something that it can start with. Now companies that are on that leading edge, again, in the process of scaling the application of agents and finance, they're frequently companies that actually, benefited from kind of early investment in data governance during the data is the new oil era, big data. But every company, especially finance functions, they too are going to have some reasonable starting place for the application of AI. And then kind of that third point, though, around, like, kinda having the wrong technology. What I think we're actually finding is that companies struggle much more with, the availability of people who can conceptualize how to apply AI to business problems. So technology is an important component. But when you think about it, right now, you have a very large number of people who are really used to, for example, interacting with AI in a chat style interface. So they have a very clear picture of what you can do in a couple of interactions. It's like upload this document, summarize it, draft an email, what have you. Kind of that next level of complexity is, hey. What what if you take a a linear process? Can you decide that you're going to orchestrate it and apply AI at specific steps? That requires a different level of capability, and it gets a smaller number of people at a typical organization. And then, you know, where larger companies, leading companies are going is in the direction of actually building swarms or networks of agents that can collaborate to ultimately solve many different processes together. That requires another set of capabilities, like the ability to take a lot of processes, envision the capabilities required to deliver all of them, and then disaggregate that into networks of agents that can work together. And, again, it's a much smaller number of people who are able to do that in a typical organization. So frequently, rather than kind of that the technology being the only thing to focus on, that, basically upskilling and building that bench of people who are able to envision how to apply AI to business problems is really important. Yep. I I really like like all of those reasons, especially your your second point just around the data quality. And it is a misconception, but it but it is certainly a challenge that companies of all stages, not just public companies, to deal with. Sigit, with this incredible blazer, mean, checkered, it's just so gorgeous. Thank you. Can you share just a little bit about how you've scaled with companies, how you've thought about data quality and and readiness, and maybe some advice that all companies should be thinking about of organizing their data in a way that they can effectively leverage some of this new technology? Absolutely. To Michael's point, data quality is extremely important. Example is that Fastly, we wanted to understand whether we could automate or we could apply AI agents for our deal flows auto management. Unfortunately, the data was so bad. It was so disaggregated. The quality of the data that was coming out was not great. And we just couldn't adopt AI for that. Are we able to adopt that in the next one to two years? Yes, we can. But there needs to be very consistent process oriented, very process and data quality needs to improve significantly there. So understanding which processes and what kind of data is coming out of that and how do you want to apply the agents or the AI tools that you're using for getting a better outcome from the products. I think that's very important for you to understand and adopt accordingly. Having that data roadmap and having a roadmap around which processes and products and data that you can apply to is very, very important before you embark on any journey around your AI products. Yep. Makes sense. And you've had to deal with many frustrated board members like myself, I'm sure, that are constantly pushing you in different directions and telling you, hey, move faster with less, but also focus on security and compliance and regulations and all that. How do you balance the demands from the board and maybe the different opinions across the executive team around some of these technology demands? Yeah, I'll take a slightly different perspective around this based on my five years of experience running the CEO org across two public companies, which are half $1,000,000,000. I've had teams of more than 50 plus. The CFO is in the spotlight every ninety days and they don't want to really mess around with anything that is working well for them. That's the conservative nature of any CFO. And the board members are based on the two or three companies that I've talked to extensively. We have had lots of conversations in the audit committees around this. There's a certain amount of skepticism around adopting and using a lot of AI products in your system. And so having that conversation around how you want to apply or how you want to adopt and how do you want to what is the ROI on the product that you want to use? Conveying that and messaging that and making sure that they understand the value contribution that the products are going to bring to you and how that is going to help the teams is going to be very important. No CFO wants to go and walk back any numbers that they give to the analysts. And that's very important for them. So any large scale changes, they wouldn't want to rip and replace any products that is not really required. And so a gradual adoption is the best way to go and making sure that everyone is coming on board is very, very important. I think one of the earlier speakers also mentioned about the people aspect. Bringing your team together and taking them on that journey is going to be critical because they are the ones who are going to apply and use the products. And so if they don't understand the value contribution that the product is going to bring them and how that's going to help them, I think you would not have as much of a success as you would expect. It cannot be a top down directive. Rather, the teams need to come back to you and tell you these are the products, these are the pain points that you have and these are the ways that we can use AI products to solve for that. Yep. I totally agree with you on just the frustrations that people have when we're constantly ripping and replacing different tools and changing what was said ninety days ago at the next board meeting. It's frustrating conversation to have. Patrick, how have you balanced this? Are there certain areas within the Finite stack that you think are maybe better suited or more ready to be able to ingest AI into, maybe month end close or some other specific areas like that? Yeah. Month end close, definitely. We heard that earlier, like some of the processes you deal with and speeding those up and using AI. It may be a gradual phasing to slowly get better there is very much opportunistic. We've been doing a lot within FP and A. I think that's, especially in public, that's a low hanging fruit for us where, yeah, what we're doing on the budgeting side and what we're presenting to the board is hypercritical. What we're doing for auditors, for any type of impairment testing, is hypercritical. But it lets us operate in a much faster environment. And I think that also helps us get, better communications with our our stakeholders where we can produce information faster for them, give them more time to review, challenge the assumptions, challenge the outputs where legacy, it's FP and A sitting in a room just trying to build up a massive Excel file, right, where that's we're trying to get away from that. I think that, you know, I mentioned we used RAMP earlier. I think that the AP has been attacked in multiple ways here. Right? That's been low hanging fruit. I think the next phase is not just, okay, that processing of information, but how does AI then help you pull that information together and tell the story? Again, I've talked about some of the regulatory, you know, compliance items that Ramp has helped me solve today just from being able to get visibility on transactions, basically, real time how they're happening. To me, the next step is how do I let AI help me stop that transaction from happening? What information is it helping me put out to the business to say, hey. We don't wanna do this or we can't do this for for whatever reason? I I think it kind of pull it all together the you know, how how we're interacting with our board. Yeah, we're not changing our strategy every 90 days. Again, maybe there's some components that we're enhancing, and that means some change there. But I think a lot it comes back to how committed to our strategy are we and being very diligent in that communication of this is where we're going. This is the road map. And again, I'm very controls focused in where we're sitting, so making sure of, hey, this is why this is still operating within those guardrails. Makes sense. Just to add to that, I think one of the advantages of adopting AI in a public company is you get tested almost every month because of stocks. And so there's an easy way for you to course correct and adapt and pivot from something to another product or another tool or another process. We used AI for doing our flux analysis as part of the month end close. Initially it was pretty patchy results, but then it got trained over time and it gave us much better output. And so that reduced our time to close and time to analyze and give outputs back to the auditors in a much more faster and more efficient manner. We used AI for doing our T and Es, and that was an easy win for us. A lot of good products out there and the adoption was very high. We failed in using AI for our Investor Relations work. And part of that was how energized are you with the AI adoption? And we tried it out, it wasn't giving the right results. And so people got frustrated and gave it up pretty quickly. So it's making sure that you're selecting the right process and the solution for your needs, I think is going to be very important and keeping that compliance effort in mind. Laika, what's the one maybe most impactful thing that you've seen finance leaders do at this stage of company? That's a great question. I think treating human in the loop as an accelerator as opposed to a perceived barrier is probably one of the best things you can do, especially early on. So right now, have a very bimodal distribution in terms of AI adoption. There is a lot of companies that are all on the far far on the far left that kind of just starting the journey Yep. And grant again, an increasing number of leading companies. But the the companies that are on the leading edge started twelve, eighteen months ago. And frequently, they would ask questions like, hey. You know, how do we actually do this in a way that allows humans to do allow us to quickly assess success? Like, how do we do this in a way where we know that we have human in the loop as a backstop, which then allows us to move that much faster in experimentation? And then also, in some cases, like, how do we do this in a way when where we our auditors can get comfortable with something that we're doing, not because we're fully relying on AI upfront, but we're actually we're ultimately relying on human in the loop review. So that's just a practical way of getting that moving faster. And what part of the organization or team do you have to loop in to be able to successfully run with a human in the loop strategy? Well, it really depends on which use cases people are starting with. So many organizations find it useful to start with, say, two to four use cases. And frequently, some companies, especially companies where it's not possible to fund the full team that's doing this, like at many smaller companies, that means you have first of all, you're spreading the load out, so it's not just one person who's bearing the brunt of, trying to AI enable the work that they're doing. Yep. But what that also means is it could be part of your finance function. It could be part of accounting. It could be from a compliance perspective. It could be from a controls perspective. Really, whatever makes the most sense for your organization. And the appropriate people are kind of naturally identified along the way. Yep. Okay. That makes sense. So maybe we have human in the loop today, but in two years, my perspective at least is that this world is gonna look very, very different for for all of our companies, for how we're leveraging AI, and and for the world as a whole. Would love to hear how you agree and disagree maybe from from all of you on on what the world looks like in in two years as as you think about your roles and and your companies. Yeah. I I mean, as John started this, right, of we're sitting here today was a a just a pipe dream two and a half years ago. Right? So, you know, I have no idea where we'll actually sit in two and a half years or two years from now. But I'm very optimistic that as we continue to move forward and and we get new technologies, more AI adoption just in general across our our ecosystem, my primary goal, again, I'm I'm looking controls, I'm also looking cost footprint. And I very much expect that AI is gonna help me kind reduce or maintain a stable cost footprint while allowing business growth. And so thus reducing my financial and fiscal impact to the business while multiplying my support that I give back to them. Yeah. When I think two years out from now, when I the main thing that comes to mind is, right now, there's this core tension between, like, to what extent will individual, vendor provided point solutions versus kind of much larger scale platforms end up solving the the types of problems that people need at scale. So the interesting thing for me is that it does feel like in two years, we'll have a much clearer answer to that question largely because people will have adopted many of these different solutions and then test them out. But that's what's exciting about that for me is that means that two years from now, we will actually have, the momentum that I think a lot of people were hoping we would have had now. Yep. And, you know, it's not gonna be it's we're not gonna be waiting two years to get those answers. And similar perspective as well, that there's a lot of AI solutions that are in the marketplace right now, and there is probably going to be sort of overload on the company, the finance teams to adopt as many as they can. And so that's going to create some stress for the companies and the teams. And so managing that expectation, managing that workload for your teams is going to be very important. And to Patrick's point, every company, big or small, is under a lot of cost pressure, particularly on the G and A line. And maintaining that flat cost neutral organization is going to be very important for CFOs and CEOs and controllers. So managing that and adopting as many products that are more relevant for you is going to be a very important decision point for companies in the next two years. To your, that's a great point about having a platform that can provide a suite of products for your needs. I think that's probably going to get we'll get a much clearer answer in the next few years. Saget, how do you stay up to date with just everything that's happening? I imagine that you have a lot of companies in your inbox that are trying to All the time. Get you to demo their products. And the answer is really just use Campfire for everything, as we all know. But how do you stay up to date with just all the changes, all the technology advancements to make sure that you're making the right decisions? I think the companies that are coming and proposing their solutions are our best way for you to learn about the technology and what's in the marketplace. And frankly, I wish Campfire was there a few years back. Native ERPs, the incumbents are not really great at building AI solutions and our teams struggle with that. It's not easy for a public company to make that decision to switch from one ERP to another. It's a very, very hard decision and takes a lot of convincing others. And so just hearing about what's happening in the marketplace, TechCrunch, the Disrupt twenty twenty five is happening right now. So that's a great place for you to know about new products. And just talking to people, this is a very good networking event where you get to know a lot of new products, different products. So that's the only way to go about it. Makes sense. I love it. Maybe last rapid fire question. What's one piece of advice you would give to up and coming finance leaders of of companies who have ambitions to go public? Be willing to accept change. Right? I mean, it's it's, again, it's gonna come at you very quickly and and yet while you wanna have you wanna have your documented strategy and kinda have that north star, right, if this is where we're heading, you're gonna get knocked off of it. Right? And and be willing to accept that, hey. You gotta sidestep or you gotta kind of, you know, pivot a little bit. Maybe not to the detriment of your changing systems routinely and changing strategies every quarter for the board. Right? But be willing to to just accept that, hey, as new things come at you, that can be a great opportunity for advancement. It's not it can maybe be slightly different from your your strategy, but lean into that change. Yeah. I would say start now. So, again, we when you think about that bimodal distribution, companies on the far right started twelve to eighteen months ago. And, you know, they they in many cases, they had several rounds of failed attempts. Right? It's there's no there's no way around it. You have to actually get the experience. You have to understand how it's going to apply in your environment. And And, of course, every company is going to look a little bit different. Yeah. Start early and bring your team along for the journey. It's a vision that you have to lay it out for them and so they know what's happening in their part of the org and how that's going to impact their world as well. And that's going to be very, very important for you. You may not have the same team structure and size two, three years down the road. And so making sure everyone understands what that roadmap looks like is gonna be very important if you wanna succeed. Amazing. Thank you. Any questions before we wrap for the Blazer crew? Great. Well, thank you so much. We really appreciate it, and and thank you to three of you for sharing insights. Thank you.