Tabs Webinar - How AI Impacts AR

Hey Sean, how are you?

Good. How are you, Ollie? Great.

I love the new Gusto background. Looks great.

Thank you very much.

A lot of inspiration got me going this afternoon. But really appreciate you making the time. You've been a great mentor to me and the company and just getting to spend some time with you to talk about all things AI and finance, I think is going to be a fruitful discussion. Before we start, just broadly speaking, excited to cover just general trends in AI, both in terms of what finance executives like yourself are thinking about and also in areas like accounts receivable where we at Tabs, which is building the first kind of purpose-built AR platform for B2B using AI to really create a ton of automation and accuracy and bring down DSO. I think it's just a very timely discussion. So very much excited to go deeper on the topic. Before we start, Sean, I don't even know where to begin, but I'm gonna at least let you introduce yourself. You've had such an amazing career as a finance executive and obviously most recently at Gusto, but would love to just like spend a minute if you wouldn't mind just giving a little bit of background on you, maybe a fun fact and kind of your mandate today.

Yeah, so my name's Sean Flynn. I'm the Chief Accounting Officer here at Gusto where I've spent the last just about 8 years of my career at this point. So, um, joined Gusto when we are a lot smaller than we are now and have had a really, really wonderful ride with a ton of opportunities. Um, lifelong, career-long at least, uh, accountant. I'm a second-generation accountant, kind of a third-generation accountant too, so, uh, the nerdiness pervades. Um, but I have, yeah, like I said, I've spent most of my career in accounting. I started my career in New York working in a few high finance institutions institutions. And then I moved to the West Coast and I took my first job in a gaming company and then I moved on to become controller at, at Blurb, which is a self-publishing platform. And then immediately before Gusto, I was with a company called Rdio, which was a music streaming platform where I was also controller. And then in the 8 years I've spent here at Gusto, I've had the, been fortunate enough to have the opportunity to build out an amazing accounting team, and then also have been involved in empowering our treasury payment operations and risk operations functions. So yeah, like I said, I've spent a lot of time with Gusto, had a real, a lot of really, really wonderful opportunities, and I'm super happy to still be here.

That's phenomenal. Really, really appreciate that overview. I think, Sean, just what a lot of our Viewers are just curious about just for someone from your vantage point, right? You've been a controller kind of in a lot of finance strategic conversations as well, kind of in the last 12 months. Before we dive into any specific use cases, just like what has been your reaction to the AI wave broadly? Like to what degree like does it come up? Has it been top of mind? Like just would love to just hear your general take on the, on the AI mania over the last 12 months.

Mania is the right word there. No, I think it's been, I would imagine our journey has been pretty similar to a lot of other people's too. And Ali, you and I had spoken about this. There was an enormous, I guess there was an enormous what felt like a breakthrough but was probably a long time in the making earlier here in 2023, late in 2022 when the consumer version of AI really came out or what we've all identified with what the consumer version of AI is. Which is, can I go to ChatGPT or can I go to Bard and can I get a meal plan for the rest of my week or plan my Italian vacation or tell me what 3 barber shops are in town? Not that that's something I'm looking for necessarily, but that was, that was to me, that was the first breakthrough of what AI was going to mean. Obviously, it was met with a huge rush of a media cycle that said here's everything that AI is going to change immediately. And if you're in this, this or this industry, you had better watch out and/or get ready and/or if you don't arm yourself with this now, you will be left behind. And to listen to some degree, I'll say it, I bought into that. There were people on my team that was saying let's find AI to hit every opportunity we can inside of accounting and risk and payment operations. Those things are still going to hold true. And you and I have spoken about this. I think the way that it's probably going to become pervasive though is going to be less with the hit you over the head and or take your job. It's going to be how can we augment ourselves? How will these softwares we already use or that we will be using augment themselves to make themselves better? Hopefully that does bring costs down and or run more efficiently as a company. So that'll happen either for softwares where using, or obviously can be applied in-house at places like Gusto as well. But I mean, in summary, I think it was an enormous wave of everything's going to change and then a sobering few months of this is how this is probably going to change. We are gonna need to let some of this settle out and actually be incorporated into people's platforms thoughtfully. And so that's kind of how I've seen it. It's a settling to a certain degree.

Super fascinating. Would you say, Sean, like early on, like, I mean, I think you said it right, which is all of us felt it like as the consumer, it was hard to not read any periodical or be on any blog or Slack group that wasn't referencing kind of the AI frenzy. After that initial wave, where would you say most of the AI, like thoughtful discussion happened internally within your finance team? Was it from Vendors reaching out either existing or new? Was it just like really smart people on the team that were kind of raising their hand and wanting to talk about it? Was it in other like, like finance groups amongst your peers? Like where did it go from like the consumer side to a more, even if it was in a slightly more reasonable and somber, slow, more slow crawl, walk tone, like where did the actual like— finance startup?

A more sober tone is how I think I'd put it. Definitely more reasonable. It's come from all angles. I mean, we as a processor of, you know, large sums of money and also large sums of information from a product perspective, our engineering teams are now constantly thinking about how to use it. There's been a department spun up internally which focuses basically on improvement through AI. So that is outside of the finance department. Now, we've also been pretty fortunate to have a couple of people within our greater CFO org who have had passion around it. And so the way I've seen it bubbling up inside of Gusto is it becomes one of the first questions you ask either in evaluating a new platform or evaluating a work stream you're working on, which is Hey, let's ask the first question. Is there an AI solution out here right now that can help us do something here? And I think the answer there to some degree, we're still in very early days. You know, we're coming down off the spike of people typing into Google, is AI going to replace my job as fill in the blank and getting more into what are the pieces that are going to plug in into my world? So I think, like I said, we've been fortunate to have a couple of people internally who are thinking about it from a cost perspective. Are thinking about it from a productivity perspective and who are pushing. And I think for the most part, I think what's been helpful to see is a lot of people move past the situation of, well, I'm not gonna let this thing do my job, but more how am I going to make this part of my life? And, you know, you and I have talked about a few use cases on where even just the GPT, ChatGPT in the early days has helped clean up either time spent communicating or getting the bulk of, bulk of some writing out of the way for you to go back in and give haircuts to. And I think that that's where we've seen a lot of it. I think, yeah. So it's not been necessarily in, particularly in the finance space. I still think we're very early days in how these are being incorporated and or sold to us.

Got it. Got it. I have a bunch of double-click questions on that, but just like As a finance leader, you, I think you already put aside like right now AI is not a necessarily a staff replacement. But what is the area when you think about like value props of AI within finance, do you think are the thematic pieces of AI that resonate highly with you that is like, okay, if it can do X or Y thematically, it is a good investment of our time and resources.

Yep. I think, I think matching any sort of data matching is going to be enormous for us. So we here at Gusto have 300,000+ customers, everybody running payroll, all of those customers running payroll, all of them having different benefit deductions taken, different pieces, obviously moving. Everybody here has gotten a pay stub. You know how many numbers are on a pay stub and how many things need to be reconciled on the back end. And then the actual money movement along with it. And that's just in our business side. So for most businesses though, there is going to be a corporate side, especially if you are highly transactional in what you do as a business, taking a set of data that comes from A over here and then matching it internal system to B, what's coming out of a vendor or a partner and doing so in ways that I think the thing that I'm encouraged about moving past is the human macro, right? And what I mean by human macro is, are you teaching a person to do a bunch of things in order in a system that the system should otherwise probably be taken care of, thereby opening that person up to do more value-add work? And again, as a large-scale money mover, we are deep, deep in, in money we moved versus money the bank said we moved versus money that a state government may be waiting for in the form of taxes. And so tying all types of that information in different formats together is one of the areas, and I'm not saying we're seeing that right now, it's one of the places that I'm really, really encouraged or hopeful we'll see is that large-scale data matching between different formats. So that's a big one for me. And then I think the other area, and I know our finance folks internally have been looking, is taking historical results, taking historical trends in either expense or revenue. And honestly, and at the risk of simplifying it too much, plugging it into AI and seeing what trends come out of it, because some of the work that I know FP&A teams do tend to focus on is understanding how expenses move throughout the year by vendor, by work stream, and having AI be able to come in and not tell you what the future is, but tell you what the past looked like and what the future might look like because of it. Would be enormously helpful.

That's phenomenal. No, I think both are things we've heard from a lot of, I think some of the smartest forward-looking finance leaders we've chatted with over the last couple of months. One is just the tremendous opportunity from a data matching standpoint or even touching a bunch of data that's unstructured and creating structure around that otherwise would not have gotten done or done very manually. It is often the least desirable work that folks want to do. And I mean, it's a lot in the theme of what we're thinking about from an accounts receivable standpoint is just the amount of opportunity AI provides on the data matching or movement unstructured data to more structured forms. So I think that's super interesting. And then on the second theme, there's so much insight and analysis. I think a lot of times people get really caught up with where AI can replace existing work, but I think folks like you are really seeing the tremendous opportunity of like things you wish you had, but you just can't get to because you don't have the internal tooling to understand that large amount of data. And otherwise, even if you had large access to data scientist teams, it would have just been, it taken too long. It's just hard to negotiate the resourcing between finance and so many other teams that are sharing those resources.

Are people short on data science resources? Did I miss that? Just a little bit.

Yeah, just a little bit. Fascinating. What, I mean, just, I mean, you and I have talked so much about the opportunity AI presents on the AR side. Is there any specific functions within finance or workflows or anything else that you think AI presents a ton of opportunity around?

I mean, the theme that I keep coming back to, and apologies if it seems vague here, but it's very much not, is there's a philosophy underlying this, which is everybody wants to do work that adds to where your company is going, right? Everybody wants to do what they perceive to be valuable work. AR is valuable work, no doubt, right? There is, there is, it is actually the lifeblood of a company, which is I can make revenue, but that revenue turning into cash is what allows me to keep going and allows allows me to recognize that revenue in the first place. I won't go down that path too far because then we get into that pervasive accounting nerdiness. But AR is 100% value add. What I've seen historically is some of the processes underpinning the parts that humans don't necessarily have to do are being done by humans. The data matching, you and I have spent a ton of time in that data matching world or the data movement world, which is or frankly just ETL, right? So a ton of ETL happening by humans, which is I used this much or our customers use this much. It doesn't necessarily match with what's over here. We need to reconcile those two things and either plan an exception process or actually go bill for it and be able to have the conversation with the customer when we're talking to them. The ETL that gets to that place is something I would pretty frankly rather not have a human have to do since the majority of the value I'm gonna get out of that person is their voice in speaking with a voice that's on the other end too, which is in the accounts payable or procurement department of a customer. And so why would I ask someone to spend time downloading matching columns, adding significant digits, taking significant digits away, VLOOKUP-ing customer names against customer IDs, anything like that. That's not, that's not where I'd like to see any people, my team, other teams spending their time. I'd like to see them talking about any variances they may have with their customers actually being able to move on receiving the cash.

Super interesting. One follow-up to that, and you mentioned earlier too that there's a lot of like global AI thinking happening at Gusto. I'm just curious for your team specifically, to the degree there's interesting AI tools, workflows, etc., are you kind of, kind of having those conversations at the team member or sub-team level? Are you trying to make team-wide decisions? Like, how are you thinking about the distribution of that? Is it just some areas you're testing it, or are you requiring more like broader like approvals and mandates just to give other finance finance leaders on the call, some guidance on what's given the amount of nebulous and wide opportunities right now that there's no like true leaders in any of the specific categories. Like what's kind of level of control and autonomy you're giving your team members to test different.

Well, I think two things I would say at the top. So I am very, very fortunate to have a team that I feel more than comfortable empowering to go search the market for what is the second part of this, which is things that solve problems, right? There are a lot of point solutions out there for a lot of different things, especially in the finance stack right now. You know, expense management being one of those worlds as an example where there's a bunch of different players, they all have their strengths and they all have the things that are going to fit your company best. And so I'm very fortunate to be able to talk to basically any team member in any part of the geography of my teams and say, hey, what is the problem you're trying to solve? And also you're free to go solve it in the way that makes sense for how your day-to-day works, as long as it fits within budget. So I think that's the philosophy we have here on decision making, at least within the accounting and finance department. The second part of that is, and, you know, I think it's gonna be dependent on how folks market AI. There's gonna be certain companies that come out and say, don't necessarily reveal that AI is happening. They're just gonna tell you that they're solving a problem. And I think that's okay. And there's gonna be other companies that say AI is a huge part of what we're doing. And listen, we're having this webinar today, right? So we know where we are and And so I think it ends up being similar to what I said a few minutes ago, which is, are you asking yourself whether this work can be done automagically? It's a boring phrase I've heard a bunch of times. Or how are you thinking about evaluating the vendors that you're using? Because everything will eventually filter into basically the value you're getting for the cost you're paying too. So if you, we have to evaluate our vendors differently and we think, well, how is this happening in the back end? How are you creating this value? And does that match up to what our contract is going to look like with you? Hopefully we'll start to see as AI jumps in, you know, the economics will be there because the team on the other side of the table won't have to staff up as large. And so a little bit, a little bit broader thinking there than probably the answer you were looking for.

But that's the perfect thing. Bob, one, one thing on it, and I think I know the answer, but I'm curious, like, have you gotten to the point where you've now built out like AI budgets, whether for your function or the team, or is it still early enough where it is insignificant and can be fit into normal like discretionary budget?

Yeah, it's, it's still very much discretionary at this point. I think that on a, on a for, that's for the group. On the company level, we're thinking about it a little bit differently, as is everyone. But I think that that is more the focus. And then, you know, I think what's, what's going to be interesting is how companies, and we're not the only company out there saying, hey, we have an AI team who is out there, because you don't have an AI team, you say, all right, we just figure out AI, right? It's going to be figure out what the practical applications of AI are, What I think will be interesting and what I would love to see happen is this, we in finance, we have a, we call them biz tech over here, but they are the team that supports our finance systems, right? So we use NetSuite and they admin and build out our NetSuite for us and same with Coupa, etc., etc. What I think would be really interesting and I hope is the future is an AI team that is functioning not wholly dissimilar from that kind of IT team that is treating our groups as customers and saying, hey, walk me through what you're doing. I don't necessarily have to, in the case of AR or even, you know, we've talked about a GL categorization. I don't need to explain to somebody in an IT AI team to, I don't need to explain, well, it's going through here, so it's actually going to hit other liabilities instead of current liabilities. And that's a problem for me. No, well, you can go further upstream. And boil that problem down to where someone at least, you know, like the AI teams tend to be engineers. Engineers happen to be very, very smart people. Anytime I've taken an engineer at something having to do with accounting, they typically latch on pretty quickly, yourself included. And so it's, we've been able to like, the hope is that that AI becomes more of a shared services team attacking problems. And applying it rather than us having to solve every problem ourselves using AI because we're not— my team's mandate is not to understand it broadly. Now it is to understand their problems and hopefully find an approach either through self-research or through a dedicated resource that can apply those problems to AI or vice versa.

I don't think we're that far from that. I think it, I mean, it's the first time I've even thought about it, but I mean, 10 years ago, if you told finance leaders like yourself that you would have like cross-functional enterprise applications teams because ERP is going to cloud or there's going to be all this data that feeds in major platforms like Coupa, you're going to need cross-function. I think no one would have believed you. And now, I mean, mature finance teams not only have enterprise apps, in many cases it sits under finance. So I love the idea that probably not in 2024, but like by 2030, a lot of teams will have a similar type of AI view, basically being a service provider to the major functions on how to use AI as a means to better, I'm not gonna say more efficient, but like there's gonna be a variety of use cases that they can really come in and define the strategy for the function.

Yep, absolutely.

One, you obviously sitting as a controller and finance generally, you also have to think about the risk side of things. What are the tummy rumbles right now as it relates to AI or just things that just feel very unknown that you're also on the other side of this being like, look, like we're just not going to be first movers here, or it's certainly top of mind and I'm kind of eating the popcorn to see what happens first before I'm comfortable to act.

Yeah, I think— listen, I mentioned we have 300,000+ customers and that's a lot of people. And we have grown, you know, we're in the, we're in the multiple tens of digits growth year over year. And so that 300,000 wasn't 300,000 last year, it was somewhere much lower. And so we're still growing, which means we're bringing on people onto the platform at a rate every year, right? There are new businesses being spun up. The small business community here in the United States is incredibly resilient, and especially over the last 4 years. And we at Gusto have certainly benefited from that. With that and with the attention that comes with Gusto, obviously we have to keep an eye on who's coming onto the platform, who is processing payroll, are they who they say they are? And so the AI world, and we've seen a lot more, you know, candidly, outside of the finance world, I've been reached out to by a lot more people in the risk world since it has a lot more to do with signals, frankly. Right. Like there are, there are signals characteristics picked up off different parts of transactions and it's through the network effect if you have the right corpus of data, I guess you could say, if you have the right corpus of data that you are about either customers or behaviors of certain customers that you're able to take different actions in the signup flow, in our case, the run payroll flow, whatever it may be. And so I think those are areas where you kind of talked about it, data science resources are limited. They are needed everywhere. BI resources, very much the same. Shout out to our data science and BI teams here at Gusto who are wonderful. We just wish there were twice as many of them always. And if you got me twice as many, I wish there was twice as many then. And so how are we going to leverage both network effects and also AI to match what some of those behaviors are and catch folks that we either don't think should be running payroll or maybe are even just worth a second look because of how they, how they acted at some part in one of our flows. So yeah, I mean, without saying too much there, it's an area where we're paying a lot of attention, as is anybody else that is moving money at the scale we are.

Super, super interesting. I know we just— Sean, I have a few minutes left. We have a lot of finance leaders, operators, just thinkers gearing up for 2024 right now.

Yeah.

In terms of predictions, right? I think we, we talked about how noisy the beginning of the year was with particularly consumer AI starting to just create a lot of talking points internally. I think generally, I agree with you, a lot of that has kind of sizzled off post the summer. But if we were just to predict going into next year, just what you're seeing now with the 2024 planning process and just generally, I mean, just given your connectivity with the, the tech space so broadly, what does 2024 in your view look like from an AI standpoint? Is it a lot of the same? Is it quieter? Is it more specialized and thoughtful? Like, what's your prediction going into 2024 and how are you thinking about it over the next couple of months?

Yeah, I mean, I think as much a prediction as it is a hope that I have is in the world of— and financial statements, I call them— they're our report card to anybody who's looking, right? And so we are audited. And, you know, a lot of companies are audited for accuracy and your financial statements are assumed to be or intended to be accurate, right? Otherwise, why would you create them? The work that goes into that accuracy and the mapping of all that and any finance leaders watching this know that at the end of any given period, the number of transactions they have sitting in the wrong place or without a title or heaven forbid in a holding account or other account, which which I think would probably make most finance and accounting people cringe. That's a world that I would love to see AI applied to quickly because I think it's a place where a lot of FP&A people spend their time, which is just getting things back, getting things back to zero would be, it would be the way I'd talk about it. Like if zero is everything is clean and functioning and able to sit where it's supposed to. If your bill for your insurance is coming through, but it's being tagged to happy hour spend or something like that. That's work that somebody then has to go do. And it would be lovely for AI to come in and say, well, like Jack's Bar and Saloon does not sound like where you would pick up cybersecurity insurance. Can we check this? Right. And so that is, that's a piece of, that's a huge amount of time that I know FP&A teams listening or watching or just working. That's where they have to spend a lot of their time. So I think that that would be, that would be what I would consider some of the lowest hanging fruit. And then I think it is again any opportunity where something is coming out of one system being manipulated in any way to match what's in another system and all get it all in the same. I, I listen, I love Excel. I've lived in Excel my entire life. I still have a PC, whatever that says about me, because the Excel shortcuts are faster and known by me specifically. But I think the ability to be really good in Excel should be, we should place a lot heavier weight on the analysis side of things than we should on getting to the place where we can finally analyze. And so cleanliness, general cleanliness is one of my, is really one of my earlier hopes here in the AI revolution of finance. So a little bit of a broom at first, and then I think we can bring the floor polisher through.

Amazing. I, first of all, I think, Sean, I'd love for us to do another one 12 months from now to just see how we got it right. What were— I'm sure there's a few things even us here on the tab side who think about this every day or think about every day are still gonna get wrong for next year. So for sure. Again this time next year. I think the thing I'm really leaving with, and especially for a lot of the folks joining us today, is to just stay calm. Like, if you're not seeing massive AI disruption in your finance workflows in 2023, that's probably right. Like, I mean, even the best-run organizations by folks like Sean with a ton of experience are still, still, it's still a crawl and a walk right now. No one is massively disrupting their, their workflows or their resourcing around AI. There's been a lot of noise around. I think there's some thoughtful places that are gonna come in the short term. But you're, Sean, if, if I were to summarize thinking about this really over the next 5 years versus the next 5, 5 months, what people thought 5 weeks, 2 quarters ago.

Yeah. And, and for me, listen, it is, you're not behind, just don't get behind. And, and that, that getting behind can feel like you could either frame it as well, oh no, there's, there's an 8-ball for me to get behind at this point. But it's really how are you going to adapt to a world that is changing behind you and how are you going to make your awareness and your usage of any of these AI skills or the companies that are coming out with them? Like it's a refresh for a lot of us in leadership who know there's a new skill set we have in front of us, which is remaining in front of something. And again, it doesn't mean that I know how to, I don't know, I can't plug into any LLM myself doing anything, but I can be aware of what problems can be solved with what technology. And that's something that is hopefully going to keep me at the top of my game.

Amazing. Well, Sean, just can't thank you enough. You've always been so generous with your time. I think you've left all of us feeling good that generally just stay in tune, but just it's going to be a little bit of a process here and we'll see a lot in the coming years and look forward to doing the check-in with you 12 months from now.

Awesome. Well, thanks for having me, Ali.

Absolutely. We'll talk soon. Take care.

Bye.