Latest / The Accounting Podcast / Beyond the Hype: Real-World AI in Accounting and Finance

Transcript

Beyond the Hype: Real-World AI in Accounting and Finance · The Accounting Podcast

• Attention: This is a machine-generated transcript. As such, there may be spelling, grammar, and accuracy errors throughout. Thank you for your understanding! Blake Oliver: [00:00:10] Hello and welcome to a special episode, a crossover edition of the Accounting Podcast and Beyond Spend, recorded live here at Emburse in Motion in Nashville. I'm Blake Oliver, CPA, and today we're diving into one of the biggest questions in finance right now. How are teams actually using AI? Not just in theory, but in practice? My co-host for this crossover episode is Adriana Carpenter, CFO of Emburse. Adriana Carpenter: [00:00:36] Adriana Blake, great to meet you. Nice to see you and happy to be here. Blake Oliver: [00:00:42] I'm really excited about this. Um, it's been fun being at the conference now for a day and talking to customers and and learning about what you all are up to. It seems like they're just like so many things going on with AI at Emburse. Adriana Carpenter: [00:00:55] So there is and you know, at Emburse, we were really incorporating AI within. I would even call it the culture of our business. So not only are we thinking about how AI is adopted for our customers, but also how we're actually adopting it internally, and it starts all the way at the top with our CEO. But then, you know, spreads throughout the entire organization. So it's really fun to really see and hear the hype and like, see how we're, you know, really adopting it internally as a company. Blake Oliver: [00:01:28] Right. As they say in Silicon Valley. Right. You got a dog food you have to. Adriana Carpenter: [00:01:32] Eat, right? Blake Oliver: [00:01:33] You have to eat the dog food. Right? Adriana Carpenter: [00:01:34] Or drink our own champagne is how we like to talk about it. Blake Oliver: [00:01:38] That's the slightly more upscale version of it. Adriana Carpenter: [00:01:40] Yeah. Blake Oliver: [00:01:41] We've also got joining us, Olga Pavlova from PizzaExpress. Olga, welcome. Great to have you here. Olga Pavlova-Grebliauske: [00:01:46] Thank you for having me. Blake Oliver: [00:01:47] So, you know, I'm curious, what is your experience so far with AI at PizzaExpress? Maybe you could tell us a little bit about the company as well. Speaker4: [00:01:58] So PCI express, we are celebrating 60th anniversary this year. So it's a big, big year for us. Um, and probably as no secret to many, at least definitely in UK, um, it is a casual dining restaurant chain that spans from Soho to Singapore, with over 350 restaurants that we look after. Wow. So lots of franchises, subsidiaries. So we we have the restaurants in Hong Kong, in Singapore, in UAE, UAE, UK and Ireland and beyond. Blake Oliver: [00:02:39] So I mean, 350 locations, you said. Speaker4: [00:02:43] And literally just in UK and Ireland. Blake Oliver: [00:02:45] And how many countries are you in? It must be. Speaker4: [00:02:50] Wouldn't be able to answer that one, but quite a few. Blake Oliver: [00:02:53] I'm just trying to imagine the complexity of managing all of that from an accounting and finance perspective. It must be a lot. Speaker4: [00:03:00] Luckily in UK we look after UK and Ireland. The accounting element. Um, also consolidation of the our subsidiaries in Hong Kong and UAE. And then we have got franchises which is done via royalties. Blake Oliver: [00:03:17] So Adriana the topic of this episode is AI in the real world, right? Like beyond what we've been hearing about what's possible, what is what is happening today with AI. So like at Emburse in accounting and finance. Could you tell us a bit about how you're approaching it in, in in your day to day. Adriana Carpenter: [00:03:42] I mean I it's evolving so quickly and it's pretty astounding to see how much it's changed just in the last call it even 12 months. And within in bursts. And within our finance team in particular, we started really using AI within our own platform. Um, it was, you know, sometimes it's a little bit hard to get that adoption going and to get people to believe that, oh, this is something I can do in my everyday. So being able to bring it in through our own platform has really helped one, I think, get the finance team to see the possibilities. Also, to prove out some of the ways in which AI can actually help drive different business outcomes. And and now it's really created an appetite. It's like getting an appetite to be able to say, okay, if I can achieve this type of maybe spend reduction because I was able to go look at, you know, what am I spending in this particular area, or what am I seeing as an anomaly then it is creating this environment where, you know, across our finance team, people are starting to raise their hand and say, there's so many other opportunities across, not just in the spend area, but really broader across that whole finance. I'll even call it the CFO org, where they can look for areas for adoption. So I think it's more you know, we started almost like a grass roots with eating our own dog food and then really expanding beyond that. Blake Oliver: [00:05:12] So give me an example. You mentioned um spend control. Yeah. How are you using it in that area. Adriana Carpenter: [00:05:18] So for us, um, recently we, you know, we just launched a travel and expense dashboard that basically is looks at your end to end travel and expense, uh, combines it all together. And we really utilize that to dig in and truly find areas where maybe our policy is, you know, we're seeing a lot of misalignment in policy. People are spending outside of a certain area. And so we've done education campaigns. We've done actually changed our policies based on some of the feedback that we're getting within that data analytics. And then one other particular area is we were looking at how much vendor spend do we get going through our reimbursement process versus through what I would call a normal procure to pay process. And we, like most of our vendors, spend to go through our, I'll call it the procure to pay because we want procurement involved. You get better rates, right? There's a lot of, um, improvements and benefits to doing that. And so by using that dashboard, we're actually targeting and finding a whole bunch of vendor spend that was occurring outside of what I would say is the preferred method of how we want to go after our spend. Blake Oliver: [00:06:32] Okay. Got it. I want to dig on I want to dig into that more with you. So so you have a you have a Po process. That's correct. You have a procurement process, but you also have spend that's happening outside of that. So I assume that would be on like corporate cards right. Adriana Carpenter: [00:06:47] So maybe somebody went out and purchased a SaaS subscription and pushed it through their expense report. Right. They put down their corporate card. Um, there's a whole bunch of reasons we don't want that to happen. We want to be able to understand what are the SaaS, you know, what are the SaaS instruments? We're we're utilizing we want to assess the security profile of any kind of SaaS, um, subscription that we're that we're enabling. We also want to make sure we're negotiating appropriate rates for that subscription. And then we also don't want 20 versions or 20 different subscriptions of the same SaaS, same SaaS solution. So by utilizing this analysis to say what vendors are literally going through an expense process that really should be going through a different process so that we can write and get the proper, you know. Blake Oliver: [00:07:35] Review, right. Adriana Carpenter: [00:07:37] It. Then we were able to then run education campaigns, go get those vendors and consolidate them over into one holistic subscription, reduce and negotiate better rates for that, etc.. Blake Oliver: [00:07:49] So you're taking the data from all of the spend that's happening. That's right. And then how are you generating this this dashboard or this analysis. Adriana Carpenter: [00:07:57] It's coming through our solution. So we had um, so we so we basically have gone to our dashboard and looked at specifically, hey, what are what kind of vendors are going through the the reimbursement process and identifying those vendors that are SaaS vendors and bringing them over and saying, okay, now let's bounce that up against we use our own invoice solution, so bounce it up against what are we processing through invoices? Where do we see matched vendors? Where do we see similar vendors. And then how do we go address that. So then we had buckets of or we had Adobe and DocuSign. And you know, do we want just one of those and not both. We had 12 licenses of Adobe over here and one enterprise over here. And so we want to merge those. That would be like an example. Blake Oliver: [00:08:43] And the analysis. How is that being done? Is that automation rules based or is there AI that's plugged in? Adriana Carpenter: [00:08:50] There's AI that's plugged into that. Because what's happening is our our AI basically is going in and reading all the invoices, all the reimbursements. It's categorizing all of those spend all of that spend information. And then it's based on that categorization and being able to visualize, okay, now show me where that spend is occurring and what category, what vendor, what what geography. I mean, in some cases, I'm looking at what departments are buying this stuff and you know what I mean? So you're really able to analyze it in multiple ways in order to, to then take action. Blake Oliver: [00:09:24] So if I've got like Adobe subscriptions happening over here on the marketing cards and maybe there's like some production, I don't know, another department that's got the same type of subscriptions going on that went through a Po. It's pulling that all together and then reporting on that. Adriana Carpenter: [00:09:40] That's right. And a lot of times What? What we find is that when it's going through reimbursement process, meaning an employee is off procuring it on their own without sort of it loses visibility. I mean, it might get tagged as software spend, but you're really kind of losing that vendor level visibility. Blake Oliver: [00:09:57] Right, right. Adriana Carpenter: [00:09:58] Without really getting to that dashboard level. Um, and so our analytics engine is able to really pull all of that data because you have all of that enriched data behind and then match things up and know that those are the same categories and then show you, okay, you know, these are the things you should go consider for taking action. Blake Oliver: [00:10:21] And I guess that's something a person could do. But it would take. Adriana Carpenter: [00:10:24] We used to do that. But it's really manual. So you'd have to try to go kind of vet and work through all your detailed expenses to try to cull that out, but there wasn't anything that was very at your fingertips. Blake Oliver: [00:10:36] So okay, so that's a great example. Uh, because I imagine like, yeah, you consolidate the billing, you get better rates, you know, per seat pricing. You really know like you can negotiate now, right? Instead of having 12 different subscriptions, you've got one. I can go to Adobe, for example, and say, hey, give me a better deal. That's right. Um, okay. That's a great example. So Olga, I'm interested to hear from your perspective, right. Um, like, are you using AI in a similar way? Is it like cost reduction, spend reduction, or like where is what's exciting for you when it comes to all this new automation potential that's available. Speaker4: [00:11:14] And a few elements to that. Um, one is the manual workload reduction, and another one is the accuracy of information that a human being would have to look into every single receipt, read what's in the receipt, be it one line, be it 20 lines on the receipt, the itemization. Whereas at the moment that's done by AI for us. And one of the elements that is very helpful is the keyword identification. So I can recognize the alcohol, the labels, the brands, etc.. But when it comes to very specific to the business, you have those very specific words that you want to find like what? And be it for likes of um, the service charges, okay tips and service charge. When the person is creating the expense report, the person is separated into the tips section because there is a different treatment on VAT. Um, sorry, on the service charges. From the that perspective. Blake Oliver: [00:12:25] I got it. Okay. Speaker4: [00:12:26] You have to separate it. There's no VAT on it and no VAT on the tips. So you have to separate it so we don't over Reimburse ourselves on the VAT. So that's tax authorities we want not to upset. Blake Oliver: [00:12:44] I imagine the employees submitting these expenses are not always doing it. Speaker4: [00:12:50] Unfortunately, yes. Unfortunately, yes. And that is expense per work is to identify and to make sure that that is done properly so that the that the tips and service charges are subtracted from the overall bill. So that bit is applicable only on that of all product items. Okay. Blake Oliver: [00:13:12] So you've automated that. Speaker4: [00:13:14] So what we did we have set it as a keyword detection within the solution. And now we get a flag to say this receipt includes this word. Have a look. So it doesn't skip it doesn't bypass anyone's eyes. It actually just stops there. Got it. It might be that everything is done correctly and no action needs to be taken. You just confirm it's okay. But in some cases you have to return it back and just say you need to separate it. Okay, that was missed. Especially if your receipt is quite long and the service charge or tips line is somewhere at the bottom. And it sort of blends in and people just forget to to separate them. Blake Oliver: [00:13:58] Well, that's great because now you don't have to review every single receipt. You can just it just surfaces the ones that need that, that extra eye on them. Speaker4: [00:14:07] Exactly. Blake Oliver: [00:14:08] Um, interesting. Adriana Carpenter: [00:14:10] Well, and you know what's coming. Speaker4: [00:14:11] I know what's coming now. Adriana Carpenter: [00:14:14] You might have been in the audience when someone gasped, when everybody gasped. So what? What Emburses, actually, um, what is coming to to the market soon is our folio. Ocr is what we call it optical character recognition. And it's basically going to be able to look at, I mean, very detailed receipts and truly go in and read it all and itemize. And it is so important because once you have that underlying data itemized within your AI engine, you can then make rules and you don't even have to make a stop and use your eyes to say, oh, there's the tip. Is it correct? It will validate whether or not it can do that now going forward on its own. And so this is where I will continue to evolve and start to be able to make those decisions for you. And it really is grounded in that data layer. So data is so critical, um, to really being able to harness the power of AI. Bad data is not going to you're going to get bad answers, right? So, um, you know, first and foremost is when you're thinking about an AI strategy is really understanding where is your data coming from? And this is part of the, the power. When we think about travel and expense holistically, it's why we keep stressing the importance of having everything together. Just like the example that I gave you where, you know, normally people think travel and expense and you're thinking reimbursable expense. If you're not looking at your expense report spend, you know, in, in conjunction with your vendor spend or your AP spend, you could miss the fact that in that one example, you have these other, you know, licenses, SAS licenses over here running through reimbursements. But and but you also have licenses over in your AP spend. So it's really important for that data layer to be really solid so that then you can start layering in more of that generative or agentic AI, which is really now going to start taking the data and making decisions on your behalf. So it is coming, and that'll make. Speaker4: [00:16:24] Even that one. Adriana Carpenter: [00:16:25] Better. Speaker4: [00:16:26] I've seen that today. Blake Oliver: [00:16:27] Well, and just how often are licenses being paid for that aren't even being used? Adriana Carpenter: [00:16:33] Oh, and that's that as well. That's right. Or somebody put it on a card and forgot about it. Maybe the person's left the company or. Well hopefully not. Hopefully you would have turned the card off and the person left the company, but. Right. But but a lot of times, like they forget about it and it just keeps going into perpetuity because nobody turned it off. Right. Or even worse, the card actually declined. Well, in that case, you're not you're avoiding the spend. But like. Blake Oliver: [00:16:55] I mean, it's happened to me in my small business, right? Where somebody, an employee leaves and we've got a bunch of licenses for them because everybody's got dozens of SaaS apps now for every business. And you miss some. Right. And they just it adds up over time. Right. You realize, oh, I've been paying for that for 12 months now and we didn't have to. So if you could somehow bring in data from, say, Microsoft or Google Workspace, right. Or your HR system and then, you know, I could cross-reference your actual list of employees against what you're paying for. Maybe identify areas to save and probably really add up for, like, some larger customers? Adriana Carpenter: [00:17:37] Absolutely, absolutely. And save and avoid. Right. It's some there might be areas to where you want to just turn it off. Blake Oliver: [00:17:44] That's right. So Olga, what is the like. What is the most painful. What is the most painful manual task that you have automated? Speaker4: [00:17:57] Just stop looking at something that doesn't need to be looked at. Blake Oliver: [00:18:03] Review. Speaker4: [00:18:04] Yes. That is definitely quite painful. We've got thousands of transactions going through, and if you look at something you've spent, even if it is a minute or two, but you spend that minute or two looking at something that you just at the end of it, you're just like, it's okay. So why did I look at it? Right. So having it automated, where all the validations, obviously based on the configuration in place, if all validation is a pass, you don't need to look at it. You literally deal with where your attention, your justification is required. Spend time where it's valid. Spend time with needed. So I think that is the key. That's definitely the key. Although at the same time you always see the bad things. So you have to occasionally to step back and have a look at those things that are good. And I personally do when I, when I have a look at the reports, if you feel like you're getting down a little bit, just change the filters and look at all these good things that are going to. So just to get that percentage, you know, look at the statistics, look at the analytics element of it and enjoy the view where this high percentage is actually going through without anyone touching it. Speaker4: [00:19:23] That is incredible. Yes. Um, another thing is, as I've mentioned, the keyword search is being able to identify if the wrong category has been selected. So when a user creates a report, they make a judgment on based on their experience, based on assumptions, on understanding. However, from the accounting perspective, when it comes to that, it's not necessarily always needs to go to where it's being pointed to. So being able to identify those records and literally to recategorize them, that's a benefit to the accuracy of personnel accuracy of which budget it was meant to be at. So when you compare budgets versus actuals, you're getting as close as possible rather than a massive overspend in one place and underspend in another place, and then trying to untangle it with thousands of records going through the ledgers. It's a task afterwards, so you actually save the time. Further down the line by addressing it at the very beginning. Blake Oliver: [00:20:32] It sounds like a small thing, but it's actually a big thing. Speaker4: [00:20:34] Right, exactly. Blake Oliver: [00:20:35] I've been there. Um, you're reviewing your variance report, and you realize something's wrong, and then you go in and you have to dig in all the transactions and reclassify recategorize. And you're doing it at the end, right? Um, so, like, tell me how, like, paint a picture for me when you say that it's, it's automatically recategorizing or is it, is it like, similar to the receipt example you gave where it's highlighting? This might be wrong. This might be, yes. Speaker4: [00:21:06] It's highlighting unfortunately not automatically recategorizing, but it's highlighting that we found this receipt in this category. Here you go. Have a look. Blake Oliver: [00:21:18] And how does it know to do that? Speaker4: [00:21:20] It's the keywords. Blake Oliver: [00:21:22] Okay. So you specify keywords. Speaker4: [00:21:24] We specify the keywords that we want to pay attention. We want to look at the flags of licenses. Just as you've mentioned with that one, we we put the an amount threshold to say have a look. If that's this account, this particular category exceeds the threshold. That might be not an Apple Store or storage for a month. But something something is bigger. Okay, so take a look because it could be a subscription. Docusign, Adobe or any other that within you know we're not unique in this area. It could be that a user subscribed to. And that's one of those items where a particular category should not exceed the cost. So then we actually can start looking into it. Adriana Carpenter: [00:22:17] So again sort of thinking about so what's next. Right. So you're right now using your keywords to identify areas to look at. Or is it in compliance with policy. Is it classified correctly because you want to make sure that when you are seeing the spend, it's you know, if it's air, it's really air is airfare or what have you. Um, we are actually rolling out a data quality dashboard for this very reason for the so that as our customers are looking at your tags, your keywords, and you see that a particular item, maybe you have gifts as a, as a category and you're finding that your data quality, every time a gift comes up, you have to it's being corrected. It's only it's only being accurately classified as a gift maybe 70% of the time. And so you're going to go in and it will allow you then to make adjustments so that you can properly match what that spend is to what the right keyword is, so that then again, you can take advantage of that AI functionality going forward to then know, all right, it's properly classified. And now I can do more with policy enhancement or policy enforcement prevention what have you. So again it gets back to data quality is critical. And so being able to actually very quickly see the you know, somebody manually went in and corrected this 30% of the time I need to go make an adjustment here so that the, the um, the accuracy vastly improves to then obviously improve your outcome. Blake Oliver: [00:23:53] And that's why the accurate OCR, the scanning of all the receipts. Adriana Carpenter: [00:23:58] It's kind of foundational, right. That becomes like the first the first, you know, sort of line of defense is you've got to be able to read everything, and then you have to be able to contextualize what you're reading. So same exact example. You may be able to read a receipt and say, oh, I can see a dollar sign, but is that a US dollar? Canadian dollar? You know what what what currency is that? And utilizing AI, you can actually put it into context. You can see that, you know, maybe there's a city in, you know, a city and country on there or a city in state. And it can infer from that. Oh, okay. This is a US dollar we're seeing versus a Canadian dollar receipt. Um, it can look at the dates differently. If it knows it's international. It knows it's, you know, day month year instead of month, date, year. Um, so it's things like that that you're really wanting to, you know, the way that AI has, has, has, has improved is it is starting to be able to contextualize the data that it's actually ingesting. Blake Oliver: [00:24:58] I just feel like there are so much opportunity for AI when it comes to spend and especially expense reimbursements, because, you know, so much of what we do is manual and it it takes a lot of accessing different systems and like you said, contextual data. But it's not that complicated, right? Like when I make my expense report, it's me going onto my calendar and looking at where where was I last month during this date range, and then looking at my card statement and trying to figure out, you know, what were these for? What event was this for? Adriana Carpenter: [00:25:33] So that's a great example. So why do you have to do that right. Well you shouldn't. So this is where um, you know, you start to get that, um, where you're bringing in all that data, you're bringing in your calendar, we're bringing in, we will go look, we'll bring in the direct cod feeds, we'll bring in the receipt, the. Blake Oliver: [00:25:53] Emails, bringing my emails even. Adriana Carpenter: [00:25:54] Yes. Bring in emails so that then you can ingest all of that data and then make associations so that now when you go into your app, it's going to say, hey, I see you traveled on May 20th and I can see from your calendar. And you had lunch, and I can see that you had lunch on your calendar. And, you. Blake Oliver: [00:26:15] Know, if you could fill out who I had lunch with and what the business purpose was. Adriana Carpenter: [00:26:18] Well, you saw that Olga and Adriana were on the calendar invite. And so it will assume that Olga and Adriana were there. And it knows that it was at noon. So it's going to assume it's lunch. It's going to know that we were in Dallas, Texas, because it says it on the receipt. So it knows it's in US dollars. It's also going to be able to separate out things like, you know, tips, tax, um, other things that. Blake Oliver: [00:26:42] Maybe alcohol versus. Adriana Carpenter: [00:26:43] Food, if that's a, you know, a particular policy that you want to enforce. And then it will create the expense report for you. So it's going to pre-populate all of that when the charge hits your credit card. You just paid for lunch. Thank you. It was really wonderful. Um, you know, you're going to you will get an immediate prompt. Hey, take a picture of your receipt. So I'm taking a picture. Guess what? You just took a picture of the receipt that only has the credit card amount and your signature in the tip. But as we know, and you just described it, we need to be able to see in finance. We need to be able to prove like sales tax, how much the tip is. You need more of that itemization of what you bought. And so immediately we're going to prompt you and say, oh, you didn't take an itemized receipt. You need to get that as well so that you're also trying to change that user behavior, right? You're informing your user, the employee, hey, this is how you follow policy. Because I'm telling you in the moment you got to get that itemized receipt. So now I'm going to take the itemized receipt. And then it's going to also bring in the card feed so that we can marry up that. Yes. The charge that we see here on your receipt is the same charge that's coming from the card company. Also, I know where you were, who you were with. And I'm going to then create my expense report and I'm going to ask you, was it internal or external? Maybe I'm going to know because in your calendar you you put the business purpose and it's going to auto populate it, show you and say, are you ready to submit? And you're like, yep, off it goes. And you haven't really done anything other than take a couple of pictures of receipts. Blake Oliver: [00:28:19] Right? It sounds like a dream, right? I mean, it's it's the we're going from I have to manually assemble this thing with a, you know, printing. I literally I, you know, we have been taping receipts to pages scanning them. Right. Making Excel sheets. And think about how much time that takes people. And these are, you know, these are highly compensated employees too. So their time is valuable, right? Like, especially like you're you're traveling sales team and all that. And instead of that, the future is it's made for me. And I just review it and submit it. That's right. Right. That's where we're headed. That's right. So how close are we to that vision? You know. Adriana Carpenter: [00:29:01] We're you know, we're quarters away, not years away. Blake Oliver: [00:29:04] Wow. Adriana Carpenter: [00:29:04] Yeah. I mean, it is evolving that quickly. And again, I think, you know, we've talked about it. It stems from, you know, also a mobile first design. Right. You need everybody is and really the, the employees that are entering the workforce today, they do everything on their phones. So it starts with that mobile first design. It is next layered with the data that that you can accumulate and then be able to interpret and then layering in sort of that, um, you know, that that AI to then make those inferences to be able to then populate that for you to review and approve. Blake Oliver: [00:29:40] So that's just quarters away. I mean, it's going to really change. Well, like Olga, like you said, it's going to change us from having to look at every transaction to being able just to look at the ones that we need to look at, need to be looked at. What would you like to do with that extra time? Where do you want to be spending your time in the future if you have all these tools. Speaker4: [00:30:06] So the first thing probably is looking at how to simplify the submission. Yes, we are still not there with having the reports submitted. So looking at how we simplify the user journey, submitting. Blake Oliver: [00:30:23] Getting them to actually do it or. Speaker4: [00:30:25] Well well, giving them the ideas of them, understanding what is needed, the purpose of it. Because I think you can get a better buy in if the end user understands the reason behind it. Doing it just for sake of doing is never going to win anything, right? But if people understand why. Blake Oliver: [00:30:47] I was sorry, I don't mean to interrupt, but I was going to say the most effective way that I've found to get people to submit their expense report is to turn off their card. If they don't. Adriana Carpenter: [00:30:56] Right. Okay. Speaker4: [00:30:57] Well, sometimes that's not the option if that card is needed to run the business, but it is giving the knowledge, giving the understanding. And yes, I can do things, but it can't do the human judgment. Something that is very specific to allocation, to the process, to the area you. It will take many years probably to to reflect human understanding and the customer behavior. Because every customer is different, be it a generation, be it a location, be it just in general, the social position customers are different and only the human being being present can make the justification. Giving people more time, taking away that burden. They actually can spend more time helping their customers, serving, providing experience that everyone is coming to see. And I'm talking from the restaurant business perspective. That's what we are there is to provide exceptional service, provide exceptional experience. You can't do that if you are tied up in the back office in front of your computer doing expenses. Yeah, you need to be out there. You need to be the face of the restaurant. So freeing up the time, we freeing up to our team members that are customer facing with finance, we actually allowing them to investigate areas like software licenses probably should not be purchased using the company credit cards. Yeah. As Adriana mentioned, we probably can get a better deal. And it's very likely we have those licenses in house already. So there was no point. You just needed to contact the right person. Um, but at the same time, it's reviewing those policies that are in place. Are they giving us anything? If it's flagged to us, what do we do about it? Do we highlight it? Do we just accept it because it is part of it? Do we actually make relevant parties aware of what what's actually going through the expenses? And do we need to take an action on it? Is it okay? Should we go somewhere else. Do we need to look into why? What's the behavior. What is driving that behavior. So it's really spending the time in assessing the information that has been provided by the by the tools and solutions in place. Blake Oliver: [00:33:28] Less time processing, more time analyzing. Adriana Carpenter: [00:33:33] And actioning. Speaker4: [00:33:34] An action, taking the action from it. Adriana Carpenter: [00:33:37] Yeah. Blake Oliver: [00:33:37] Getting stuff done. Actually making things happen. Yeah. Right. Adriana Carpenter: [00:33:41] Right. Blake Oliver: [00:33:42] Yeah, that's that's what matters in the end, I suppose. Yeah. Interesting. I'm curious. Um, I mean, I think it's I don't know if we actually said it explicitly. It's implied, uh, that, uh, you know, your company uses in verse. I'm curious to know what other solutions are part of your automation tech stack along with Emburse. What's your ERP. Speaker4: [00:34:04] Oracle JD Edwards. Blake Oliver: [00:34:06] Okay. And do you integrate that with Emburse? Speaker4: [00:34:08] Yes we do. Blake Oliver: [00:34:08] Okay. So then it just everything from Emburse flows in and like how does it how does that work. Speaker4: [00:34:15] The magic just happens. Blake Oliver: [00:34:16] The magic just happens. There you go. I guess that's it's funny because we call that that's sort of like old school automation now right. But that's, that's that's pretty new if you think about it. Right. We used to have to enter all that manually. Yes. Speaker4: [00:34:27] But nothing. Blake Oliver: [00:34:28] I know. Adriana Carpenter: [00:34:29] It's amazing. Blake Oliver: [00:34:30] So you're going to have Emburse you know, Emburse scans everything uh analyzes it all, gets those reports automatically generated. Right. Um, approvals and whatnot and reimbursements and pushes all that into the ERP for you. Yeah. Um, now, did you have to do it the old way? Have you had experience? You know, the old way. And then okay, so of course, you know, for, for all the, the, the young folks listening quantify, can you quantify for me the difference in terms of like the time it takes now versus what it used to be? Speaker4: [00:35:05] I wouldn't probably be able anymore. We've introduced Emburse solutions in 2018 with expanse and then progressed every year a module brought in into the business, um, trying to remember how long it used to take me to do all expenses. I know that it was clunky. Blake Oliver: [00:35:27] Perhaps in terms of like, headcount. Um, you know, like like how many people did it take to do it, or were you the one having to do it? Speaker4: [00:35:37] I was one of those who had to do it. I did, I did I used to work in accounts payable, so I used to have to deal with a two types of invoices, the EDI ones, the electronic invoices, We absolutely loved it and it was like a dream life. Blake Oliver: [00:35:56] It just flows so easy. Speaker4: [00:35:58] That would be amazing. Um, so massive volumes going through EDI and then the rest of the invoices, because not all suppliers have got those capabilities. Blake Oliver: [00:36:09] We have to keep them in. Yeah. So you started in accounts payable? Speaker4: [00:36:13] I did start in accounts payable. Yes. So I've seen that both sides of it being that end user who actually keys in sends the emails asking for authorizations, chasing because he doesn't have a response. The invoice files sitting in the trays on the desk. So yes, I've done all of those ones. Blake Oliver: [00:36:32] I started as a bookkeeper, and so I always feel a connection with anyone who actually started, you know, at the bottom of the rung, the ladder, because we understand the pain of those of those. Yes, of anyone who has to do data entry. Um, it's amazing to think like I, you know, I started maybe 20 years ago as a, as a bookkeeper and, and in 20 years we have gone from a world in which we key in transactions on the keyboard with a number pad to all of it just flows in right automatically. And being able to create rules to automatically categorize transactions, to generate reports. Um, and now we're going to take it to the next level, which is I only have to look at 10% of the transactions now or, or whatever doesn't get the confidence score that I determine so well. Adriana Carpenter: [00:37:27] And if you take it even a step further, you know, what is the future look like? Think of what I just described with the expense report where, you know, it's looking at your calendar. It's looking at what, um, you know, where you were. It's combining all of these data pieces in order to create your expense report. Part of the reason that's so powerful is because you're helping the employee be compliant, whether while they're while they're doing their travel, while they're doing whatever they're doing. Think of it another way as you start to think beyond just travel and reimbursable expense and you start to think about your vendor, spend, sort of your AP side. It's hard for employees to know how to do, you know, what is the rule of how do I go buy something? How do I go buy something? Do I need a Po? Do I need to talk to my finance person? Like, what's the process and what am I allowed to buy something with one level of approval? Two levels of approval? Why does it have to be so complicated for the employee? Blake Oliver: [00:38:28] Right. Adriana Carpenter: [00:38:29] So ultimately, you know, we see a world where you're putting AI in one place for an employee to go. You go to your Emburse app and they're going to go in there and say, what do you want to do today? I'm going to book a trip. Okay. Tell me where you want to go. Tell me what it's for. Tell me you know where you want to stay. And again, combining sort of that agentic I along with. Which means it's going to be querying you and asking you, what do you want to do next and how do you want to get there? And then it's going to be assessing, well, what are the policies and what are the rules of the company. And let me give you as an employee, well, if you're going to go on a trip, I'm going to only show you what is within policy. And if you wanted to say, well, I actually want to upgrade and I want to go outside of policy, then I'm, you know, maybe I've told them, okay, that's fine, but you're going to have to pay for that yourself, right? So, so it's going to help the employee be compliant. But it should also be able to do that on the vendor side. Oh I want to buy a SaaS subscription. Okay. For that you need to get approval from security, get approval from you know, set up a Po and it will just guide you through the process, as opposed to an employee trying to have to go and pull down a policy or try to go look for, you know, how do I do this? And you know, who am I emailing or slacking to say, where do I go? Um, it's the same idea and it's across all spend. This is why it's become so valuable to really have sort of this, you know, central solution or central, um, I guess it's a central solution, but a central process and sort of eye over all of your spend because it really helps streamline that whole process. Blake Oliver: [00:40:18] It's like, um, and when you talk about that vision, it's it's like, it's not that in births is just executing the policy anymore. That's right. It's that it is the policy. Right? It's the I mean, it the agent that you chat with knows the policy. It knows who to go to and to. The approvers are. And so, um, I don't have to read that. I don't have to go. Which employees? You know, I'm a bad. Maybe I'm just a bad employee, but I would never read any of the, uh, any of the. Adriana Carpenter: [00:40:53] You ask me where the policy is, I'm like, it's right there on the front screen. Like, right, every time you go in, you just click right there on that link. You're like, oh, okay. Blake Oliver: [00:41:03] Yeah. But you know, it's like people are busy. And then especially, you know, your high performers in particular can't be bothered. Right. They're just I just want to do my job right. And um, and so then they go outside of it because they're just like, I need to get this done. I don't have time to deal with this. That's right. But if you make the app, uh, able to navigate all that for them, I just chat, I chat with it, and it helps route me to who I need to go to and get the approvals I need to get and do it the right way. That's right. So that's a load off. Adriana Carpenter: [00:41:30] And this is what you're talking about, that that's the beauty of where AI is leading us. It is to really free up both the employee who is trying to spend in some way, and they want to do it the right way, they just need to know how. And it's allowing them to do that in a very easy manner so that they can actually focus on customer service and, you know, in doing what they should be doing. And from a finance perspective, you know, we have a lot more control over how employees are executing because we're leading, we're leading, we're leading the witness. We're basically showing them this is how you do it in order to be compliant. And so it's very preventative. You know, in accounting and finance we love preventative controls because why it prevents the bad behavior. It prevents the negative from happening. Blake Oliver: [00:42:20] And so if those controls are too hard to work with you end up preventing the good stuff. Adriana Carpenter: [00:42:25] That's right. And so there is a balance. Right. So that's exactly right. Blake Oliver: [00:42:29] You want to you don't want to prevent people from doing their job. Yeah. Well that's why like often like finance like in a in a in a bad situation. Right? Like finance is perceived as being negative. You can't do this. You can't do that. It doesn't fit into the budget. And this helps us not be that we can be enablers. Um, control spend without like, everybody feeling like they're micromanaged. I think is the dream. Adriana Carpenter: [00:42:58] And then the finance professional, you know, they're now working with the person in the business to say, well, why are you trying to spend that money? And how can I help you maybe prove you know, what is the ROI you're going to get from that spend? Why why are why do you want to do it? What is the business outcome you're seeking, right? So the finance professionals spending time on that instead of spending time. Okay, well you need to fill out this Po form and then you need to like go, right. And instead of sort of shepherding a process, they're really now, as you described, like helping to be a partner to that business person. Blake Oliver: [00:43:31] Like you said, the ah, like what you said about the ROI, right, is there's a reason that this employee wants to spend the money. Help! Help me prove it out. That's right. Me? Yeah. Adriana Carpenter: [00:43:43] Help me articulate to the business. To the decision maker, whoever that might be in your company, that it's. You know, we should spend this money because this is what I'm going to get in return. And too often, the finance team never even gets to that point because they're busy doing a bunch of. Blake Oliver: [00:44:01] Closing the books. Adriana Carpenter: [00:44:02] And so you've got to be able to free that team up to then do what's really more meaningful. So it works on both sides. Blake Oliver: [00:44:11] So, Olga, as we approach the end of our session today, um, do you have any advice for finance leaders who are perhaps looking to adopt automation AI but aren't quite there yet? You started in 2018 with Emburse, but I think you probably that journey began even before then. For you, I You, I imagine, with automation. Speaker4: [00:44:36] With automation? Yes. We had an element in finance world since. Since I've joined PizzaExpress. Blake Oliver: [00:44:43] So what advice do you have for finance leaders who are overwhelmed by this? You know who are haven't done anything yet. They haven't started. Where would you know? Where would you say they should? Speaker4: [00:44:54] I think the very first one is not to be scared of automations. There's been so many times mentioned today the trust needs to be built. Okay, but you have to make that first step. It's not to be scared. It's not to be afraid that it's not going to work. It's going to ruin everything. It's, you know, processes will fall apart. No it won't. So the very first one is to see probably where is the most of the time spent. What are repetitive tasks that are being done. Because it's those repetitive tasks are the opportunities. If one person can do five people can do, ten people can do. Maybe it's time to automate it and let those one, 5 or 10 to actually bring the best elements in what they actually the best at, be it analyzing, be it recommending, being, finding the new ways of doing things. But it's not to be afraid of what automations not to be afraid of technology. Blake Oliver: [00:46:01] So you're not worried it's going to take your job? Speaker4: [00:46:03] I wouldn't think so. If it was to take the job. I think I might have been out of job already because I'm the one looking after the system. If everything's automated, that's it. But it's not. The world is changing. And if you've put an automation five years ago, it's not necessarily to to say that that automation will be the best way forward two, 3 or 5 five years later, it might need to change because the integration is between two systems. It's not that one system is evolving. All systems are growing, all become better, all become more technology savvy. If it was always about flat files before, now you're looking at rest APIs and you've got connectors between the systems. And both have got a connector now. And every company, every solution is changing. So that means you have to revisit it. You'll have to look. Is it more beneficial to do it better to do it differently at the same time? It will give the opportunity to review what we've been doing yesterday. Is it the best way to do tomorrow? Reassess. Challenge yourself. Challenge what you've been doing earlier this morning. Last year. Last month. Last week. Just challenging reviewing. It must be a better way. And in most cases, the answer will be a yes. And that's where the value will be added. Do it in a better way. Now spend a little bit of time to save time in the future. Blake Oliver: [00:47:44] Spend time today to save time in the future. Yeah. And that and it's clearly worked out for you. Speaker4: [00:47:51] It has. Blake Oliver: [00:47:52] Yeah. And like, I like what you said about you run the systems. Right. So you've you've created these systems to automate this work that you used to do yourself manually. Speaker4: [00:48:03] And now you haven't developed them myself but I have specified implemented them. Blake Oliver: [00:48:08] And you you oversee those systems now. And there will always be a human who has to, you know, make sure everything's working properly. Right. So, Adriana, how about from your perspective, advice for finance leaders looking to You implement automation? Speaker5: [00:48:29] I would say first. Adriana Carpenter: [00:48:31] Educate. Speaker5: [00:48:31] Yourself, educate yourself, educate your team. Adriana Carpenter: [00:48:35] Um, I was actually just, you know, talking with our CTO and CISO last night and talking about creating specialized education just for the finance team on other ways or more ways that we can look and adopt AI. So I think education is really important. The other thing I would say is. Blake Oliver: [00:48:57] And if I could just interrupt you, where do you go for that? Where do you look for education around this? Speaker5: [00:49:04] I mean, I. Adriana Carpenter: [00:49:06] I look I go to a lot of different, um, we have like a CFO organization that I'm a part of, and I get ideas from that. I get ideas from others in, in the industry. I get ideas from my CTO. I just asked him last night, hey, these are some ideas that I have. What are some other good classes that you're aware of? Right? Go to other experts in that field. Talk to your data team. Um, so I think there's a lot of different areas that you can go to. I tend to sort of lean towards. Yes. What can I go search for? What are the communities I'm a part of that are sort of feeding that information to me. And then who are people that I know that are I would call much more experts at it than I am. And what can they recommend? So those would be the areas I would go to. And then I think the other thing is, is, you know, as a finance professional, there are a lot of solutions out there that are available to to assist, whether it be in, you know, not just in spend management but cashflow forecasting or, um, just even just your normal forecasting and, um, around payments. We didn't even get to to that topic. And how you think about, um, streamlining and optimizing sort of treasury. So there's a lot of different areas and a lot of solutions. Look for partners that are investing in leading in these areas because they can also make it easier as a finance org to adopt and then continue to iterate. I mean, you just said technology is constantly changing. And so if you can find partners that, you know, are also continuing to invest and change and adapt as the technology is adapting, it can ease your burden because you don't want to be doing that yourself. I mean, you could, but, you know, it's easier to leverage the partners and that's in that situation. Blake Oliver: [00:50:51] Well, and the tech is changing so rapidly that, uh, if you implement a solution that's built for now, then you might have to switch it in 5 or 10 years or less even. So if, like you said, if you want, if you want to benefit from what AI is going to do in accounting and finance, partner with a vendor, a developer that is is building that into the platform that has a vision for it, like the vision that you painted was very appealing. Adriana Carpenter: [00:51:25] I can't wait for it. Blake Oliver: [00:51:28] So, you know, that's that's the benefit to of, um, you know, when, when the finance teams, uh, at software companies are using the software themselves, um, then they, you're basically, you know, like the biggest advocate for accountants at Emburse, right? That's right. Because you have access direct access to that, that team that's building it. Speaker5: [00:51:51] That's right. Blake Oliver: [00:51:52] Yeah. I'd love to see that. Well, thank you so much for this conversation. It's been really wonderful talking to you both. Olga. Adriana, um, is there anything else you'd like to say before we close out today? Adriana Carpenter: [00:52:05] I would just say, don't be afraid and jump in. I think you actually led with that at the top saying you really need to embrace it. You know, it is AI. It's not just the future, it's today. Like you need. Don't wait. Get started. There's no step that's too small. Um. And use others in your community to help you. If you're nervous about that first step, you know, and if you're an accounting director in a particular org, go talk to some other accounting directors and ask them what they're doing, like find others that can also help lead lead that, um, or blaze that path with you. So it's always helpful. Blake Oliver: [00:52:43] Uh, one thing that I do a lot and I don't know, maybe this is because, uh, I don't have enough friends is I talk a lot to. I. Adriana Carpenter: [00:52:50] Um, you can also chat gpt the question, I guess. Blake Oliver: [00:52:53] Right. It's it's. Well, it's great to have like, groups of professionals that you talk to, but also it's kind of incredible what you can learn just from, uh, these tools like ChatGPT, obviously everyone knows, but there's also perplexity. Yeah. Um, and it's, it's it's like I remember when Google became a thing, right? We could Google for information. And the world's information is now at your fingertips. This feels like a ten x or 100 x type of situation where now, um, you can if you're curious, you can learn almost anything. Speaker4: [00:53:29] So I and I guess my tip would be, is don't be afraid to speak to people about what do they do in the business. Blake Oliver: [00:53:38] Um. Speaker4: [00:53:38] And it doesn't have to be just that solution. You literally can go to the conference as to the events organized by various businesses, and your topic might drift away from the original purpose of the conference, but you'll gather the insight into what others are doing, what other issues they face, what have they done differently, implemented something, and then speaking to your subject matter experts within the business to understand where the pain points are, where the time saving would help them. Because it's not for myself. Probably not to you to decide what to do. It is it all comes from those who actually are doing those processes. They are the key people to give you the information, to take the next steps to enhance the processes. Blake Oliver: [00:54:38] I think that's great. So it's talk to your professional colleagues, get out there with fellow CFOs and controllers, and get out of your office and talk to the stakeholders in the business. Right. Um, I think it's get out from behind the screen is kind of my takeaway from that. Wonderful. Well, Adriana, Olga, thank you so much for your time today and, uh, hope to see you around soon. Adriana Carpenter: [00:55:01] Thank you. Speaker4: [00:55:02] Thank you for having.