Why Sage Is Building Its Own AI Instead of Just Plugging In an LLM
There may be errors in spelling, grammar, and accuracy in this machine-generated transcript.
Doug Lewis: Welcome back. The most official unofficial Sage Intacct podcast in the world. Every episode, we bring together Sage insiders, industry leaders, global voices, half of them deep inside the Sage ecosystem, half of them bringing a fresh perspective well outside of it, for honest conversations about accounting, finance, technology, and [00:00:30] love that you won't find anywhere else in the world. I'm Doug Lewis.
Matt Lescault: I'm Matt Lescualt.
Doug Lewis: Our guest today, an actual expert. It's a it's a rare one on this podcast when we get a real expert in the room. Taylor Brown CPAMPA whose career is centered around helping organizations work smarter through technology, started a career at Amazon, where she developed a passion for process improvement and operational efficiency before moving into accounting and finance roles, where she experienced firsthand how technology can transform [00:01:00] the way teams work. The passion eventually led her to Sage Intacct, where she spent four years as both a senior solutions consultant and account executive, helping organizations modernize their finance operations. Today, she's a director and digital practice helping organizations evaluate technology, improve business processes, and drive successful digital transformation initiatives. That one was a mouthful. She's also an adjunct professor at the University of Texas, Austin, [00:01:30] where she teaches both undergraduate and graduate accounting students at one of the nation's top accounting programs, which I just learned from Taylor that it's one of the top accounting programs. Taylor, welcome to the show. How are you?
Taylor Brown: Thank you for having me. I'm doing well this morning. How are you?
Doug Lewis: Oh, I am doing excellent. And I have to ask you since since you are a professor, it's rare because everything I've heard my whole life is those who can't do teach. Now, is there any truth to that? I have to know.
Taylor Brown: I guess there [00:02:00] is some truth to that. I mean, I feel like even in my director, my consulting role, I'm not actually doing like journal entry recording and preparing and walking through closed processes. So to an extent, I'm teaching and helping others with efficiency from the lens of understanding what takes place. But I'm not in the action as a player, I'd say.
Doug Lewis: See, but, but you're kind of you're in the gray area where you have actually done it. Yes, [00:02:30] you have practiced what you preach and now you preach it. Well. And apparently one of the top accounting programs in the country, which I didn't know.
Matt Lescault: I don't think she has a green visor or a pocket protector though. Just saying.
Taylor Brown: I feel like I'm like one of the players on a team that played professionally for a couple of years and then just realized I was more passionate about coaching.
Doug Lewis: So you did. So now I gotta know because this is not part of our our discussion today, but I just have to know from my own personal knowledge of all of the professors that you work with [00:03:00] in the higher education system, which I imagine have been quite a few over the years. Does that have any truth to it? Those who can't do teach?
Taylor Brown: I think so. I think there's a different mindset with academics, especially tenured staff, that's more research based. Um, even part of the reason they brought me in was because I was more in practice, or at least closer to the action, just to kind of help students have that real world exposure. And they also brought me in because I wasn't the classic just public accounting [00:03:30] route. So a lot of the UT kids obviously go Big Four. They go to firms, they do either insurance or tax. But as I think the job market gets more competitive, it's important for accounting majors to understand that there's different things that you can do besides auditing and tax stuff, which obviously is very important. Um, so I'd say there is some merit to that, especially more for the research based faculty. But I also think the research based faculty a lot of times is on different accounting [00:04:00] standard boards. So a lot of them are on the FASB at UST. Some of them are on the Gasb. So they're helping to set the standards that are used in practice or they're doing side consulting projects. But many of them are like me and maybe spent some years doing classic accounting work and then went back and either got their PhD or master's degree and decided to teach.
Matt Lescault: And don't let her undersell herself. When she worked at Sage, I worked with her and she was the only sales rep that if she said, here's what is needed on the SOW, [00:04:30] I didn't even question it. I was like, gotcha, I get it back to you in ten minutes. We're good to go. Um, so don't let her undersell her knowledge and actually ability to deliver.
Doug Lewis: That's some high praise coming from you because all this guy ever does is tell me how terrible everybody is at everything, including myself. But, you know, that's.
Matt Lescault: No, no, no, that that's just you. I've never said anything about anybody.
Doug Lewis: I just experienced that small piece of of slice of pie of Matt. That's great. That is. That is excellent. Well, Taylor, thanks for joining us today. We're bringing you on. So [00:05:00] release notes the exciting, sexy world of release notes. Everybody loves them. Everybody loves them. Now, the challenge we've run into in the past with release notes is both myself and Matt. We're we're clueless. We're buffoons we don't know anything about. We don't understand it. But more importantly, we don't understand how it impacts those inside the Sage ecosystem. So broke it down to four categories. We'd love to get your off the top of your head. Thoughts on each one? Kick it off, Matt. Category one. Why don't you why don't you. Why don't you dive [00:05:30] into this one?
Matt Lescault: So I guess the question that we really are asking here, Taylor, is when you look at how finance teams actually manage cash day to day, where does the most time get lost and does this release address the right problems?
Taylor Brown: I think it definitely is helping to address how cash is being tracked more effectively. Um, there's always a ways to go. I think there's different areas that we [00:06:00] can improve that type of tracking taking place. And you have to look at AP processes. You have to look at the AR process. And then you have to look at to me the reporting of that information. So I think this release addresses some of the apps, making sure we have more reconciliations around things like three way matching. It gets more intimidating with, of course, AR based tracking as well. So we really understand more at the cash activity of an organization. Um, when I was reading through [00:06:30] the releases, the first thing that came to my mind was the cash conversion cycle. And we talk about that sometimes in class with my students. And it really gets into what, how does business activity convert into cash effectively and what is a timeline look like for that? And so I do think the release items that have come out are addressing really that AR turnover and that AP turnover piece and then expediting that time to having cash inflow or having cash outflow or getting visibility. If you need to follow up with vendors faster or [00:07:00] follow up with customers.
Matt Lescault: So let me ask this question. I mean, there was matching before, okay. Before this release, what really changed about this release when it came to that three way match? And that's impactful for these finance teams.
Taylor Brown: Yeah. So I think it's a bit of a more effective matching. I almost hate to say that with the three way match before it was testing, um, basically different thresholds from a matching standpoint. And so even though [00:07:30] it was doing in, in essence three way match, it wasn't a full three way match process. And so I think this moves us closer to hitting that point. If you had organizations that were truly doing three match processes, it was still not quite effective for them in terms of having that full reconciliation between all of the different transaction records that they needed to make sure we're matching. And so there was some more work, I think, for the product team to do to get us there. And this release is, I think, moving the needle a bit more in that direction to make Sage, I think, [00:08:00] more competitive with organizations like NetSuite that have more classic three way match programs.
Matt Lescault: And we go back to your statement around, uh, your teaching at the university and what your, what you teach the students there. Why is this type of thing so impactful for the finance operations? What, what do you really educate around that really change management, if you want to put it that way?
Taylor Brown: Yeah. And I guess just to make sure I'm understanding when you say me [00:08:30] educating more around the change management, do you mean helping them to understand the importance of systems or understanding the importance of our AP and AR processes?
Matt Lescault: I think it's the latter. Just since we're in this topic around the a PAR of how that has such or can have such an impact and what, what, what do you really talk through in that when you're, when you're teaching?
Taylor Brown: Yeah. So we talk about, I guess we'll start from more of an AR standpoint, the importance of managing customer relationships. And I like to [00:09:00] start there because I think a lot of times you think of accounting as just booking journal entries and tracking transactional information, but the accounting team really helps sales, make sure that they are receiving the dollars that they're supposed to from a sales cycle standpoint. So there's a lot of interaction there, especially for smaller organizations. And as Sage targets, you know, SMB and mid-market organizations, having that visibility quickly allows accounting then to go to the sales team and say, hey, that's great that we have these contracts outstanding, but we're [00:09:30] not receiving from these few customers. We need your help following up on them from a collection standpoint. And I think it also helps with notification processes. So a lot of times, part of what gets kind of lost in the swing of things, it depends on how many collections people we have on our team. So who can actually call, who feels comfortable calling, following up and making sure we're actually receiving cash for the revenue that we're earning. And so we talk a lot in class about the relationship between revenue, which is obviously accrual based, but also the importance of actually [00:10:00] receiving the cash itself. Um, I start out class and I always say cash is king. At the end of the day, I can recognize all these revenues, I can recognize all these expenditures, but if I'm not receiving the cash on that, and if I'm not paying my vendors, I'm going to go out of business. And so I think we have to keep that central when we think about the receivable process, is that the goal is to have a cash inflow and to have a predictable cash inflow, and that starts with having great customers, understanding credit limits, being [00:10:30] able to follow up with those customers and notifications if they're overdue bills or enforce penalties on those types of things. Um.
Doug Lewis: A piece to that you tossed in there is the ability to kind of follow through follow up on these, you know, items. If you have the data. Everything I'm hearing about younger generations is they're desperately just deathly afraid to pick up a phone and talk to somebody. Call someone you know, are you seeing that? And in kind of the student population over the years, what?
Taylor Brown: Yeah. Um, it's been very interesting. I think the [00:11:00] class that was in school during Covid, definitely a lot more trepidation, I think around in-person interactions. I mean, we were doing class online. Um, I do think the new generation has gotten a bit better, but even from just an office hours standpoint, I was office hours virtually and I host them in person. And I don't even get that many students that come virtually, which I'm trying to imagine being in school and being able to like log into my computer after I like, wake up in the morning and go [00:11:30] to office hours and not going. Like that's confusing for me. And so I do think there is a propensity for them to kind of sit back and almost like wait to receive information versus like seeking it out. Um, I do think we try our best to really talk to them about the importance of communication. I try not to like get on my soapbox constantly, but I'm like, y'all, at the end of the day, you can be a very effective person, but if you don't have good relationships with the people that [00:12:00] you work with, eventually that's going to create pain points for you with progressing forward. And so you get a lot of really smart kids that I think don't quite understand that. And it'll be interesting for me, I think, to see how that continues to impact, um, the workforce as we go forward.
Matt Lescault: And one of the things that I want to do is just sort of like highlight something I think is so important about what you said previously on the, on, uh, on the payable side and receivables, the cash side of things. I think historically, and [00:12:30] I hope I say this every episode, I think historically businesses and business leaders have looked at the finance function too much as a cost center and not as a profit center. And I think that these type of capabilities and this type of mindset is how finance teams deploy that and say, hey, if we're getting cash in faster and we're putting cash out slower than we are now creating interest income, investable cash for the organization, we're creating bottom line reality to the organization. And I think that's really important.
Taylor Brown: Yes, exactly. [00:13:00] We can focus more on capital projects. We can focus more on expanding the business. I like to say it's taking you from a tactical finance and accounting team to more of a strategic organization. I mean, I think that's to me, the whole pitch of getting better technology is not that it's going to steal someone's job, but it's going to make someone's job more interesting. You're going to be able to work on. To me, the types of projects that are more fulfilling. And I, you know, I'm biased, but I feel like accountants are some of the [00:13:30] smartest people in the organization and to an extent, working on very transactional processing and recording data, I think is underutilizing those professionals.
Matt Lescault: Well, I think you just gave us a segue, Doug, to our next topic.
Doug Lewis: Yeah, which is phenomenal. I had, I had another question based on what you said. But you know what? I'll I'll spare everybody the boring features. So artificial intelligence, AI, the robots are coming for us. There's no question about it. It's just it's a when, not if. At this point, I think we can all agree on that. We've all seen the movies. It is what it is. So [00:14:00] another part of the release notes, the fun part. Copilot. Finance intelligent agent. Oh boy. Where should we start here, Matt?
Matt Lescault: Well, I think the, the, the question is like, how often do finance teams actually get timely answers to the questions that matter? And what does it mean if anyone in the org can just ask the ERP directly, how can that impact the organization?
Taylor Brown: It's going to impact significantly. But I think [00:14:30] we always have to keep in mind that it's only as effective as the people that are using the tool. So I do a lot of sessions in the past, you know, 6 to 9 months about AI. And I think people go in expecting to just hear about what the technology does, but I like to spend a lot more time talking about the people and the processes and the governance around it, because what we're seeing is that people are getting AI tools and then they're not using them effectively. And then they say, oh, I didn't see ROI. [00:15:00] And it's like, okay, well, you didn't have a plan. You didn't have.
Doug Lewis: I gotta jump in right there. So can you, can you think of like a real life use case where you saw that take place where it's like, hey, we invested in it. It didn't work because we used it the wrong way. You know, is there a specific example that comes to mind or is that more of a, you know, just kind of a general rule of thumb? You do see?
Taylor Brown: So I see it happen sometimes even with AP automation, where teams don't realize that the more frequently you use that functionality, the smarter it gets. So you'll [00:15:30] have individuals come in and use it a few times and say, oh, it's not recording the information for my PDFs appropriately and just stop. And it's like, hey, AI is like having an intern. You have to train that intern, and the intern gets better over time, and then maybe you bring them in as an associate. So you can't just use AI for a week and then say, you know what, that doesn't work. I'm going back to my old ways. So that's an example that I've seen come up actually a few times, um, which I think is a coaching thing as well [00:16:00] for us as implementers to make it clear to people that AI takes practice and it takes interaction to be able to use it effectively.
Matt Lescault: All right, so jump into it, tell us what Sage is doing with their copilot and tell us how that sort of either strays from or is in parallel with, you know, your thought process.
Taylor Brown: Yeah. So I think Sage's AI strategy is interesting. And I think we're going to talk a bit about the MCP [00:16:30] later, but they basically are working on internal AI development and then creating an ecosystem that allows you to build your own agents on top of sage or connect other tools to the system. So the Sage copilot functionality, which I always say is not Microsoft Copilot, it's just named similarly to be confusing, um, is the main area that they're building on. So even with the financial intelligent agent, you access it through that little copilot chat that you [00:17:00] have today for the close assistant or the close workspace, or what they're calling now, the close agent in the tool. Um, I think they're focused on helping with controls around processes and with visibility. So you see with the workspace, we're getting the collaboration where people can see where they're at from a workflow standpoint, which you had with the checklist, but it expands on that with having more interaction with being able to go back to the initial workflow you were in, as well as sending out actual notifications [00:17:30] about where you're at or when things are due, which I think alleviates some of the micromanaging that might have come with the past checklist functionality. Um, the close assistant to me, I always joke it's like the paperclip that used to be in Microsoft Word. You should just have it open as you're working through processes. I think you're looking for Clippy.
Doug Lewis: I'm pretty sure it was Clippy. Clippy.
Taylor Brown: Okay. Yes. I'm always like that. Probably dates me a little bit, but I remember thinking he was so fun and he'd unravel and move around the screen. So maybe we'll get [00:18:00] there. Um, and then you have the variance analysis functionality as well, which is helping you to understand obviously differences between um, budget versus actual on a location in a department basis, but the financial intelligent agent is getting into, I think a more, um, I don't want to say it's more expansive, but having interactions with the AI that allow you to prompt it more freely. So historically you had to say run close assistant, and then it's going to pop that out [00:18:30] for you or say, run a variance analysis or show a subledger report with the new financial intelligent agent. You can use different framings or promptings to have it pull information for you. And then the goal in the long run is to create these workspaces or these really interactive dashboard views. And I use that term lightly because obviously people might say the dashboards are interactive now, interactive now, and that you can drill into transaction records. But the thought process in the long run with the financial intelligent agent is that you're actually doing [00:19:00] all of your workflow processes through those workspaces. So a good example is if I was looking at AP now, I could use the financial intelligent agent to actually help me prioritize who to pay. It's not just going to give me a list of bills, it's going to tell me why I should prioritize certain bills, even focusing on certain vendors and saying, you know, we have a big chunk of money outstanding with this vendor. Maybe you should pay them first from a vendor relationship standpoint. And then if you get into the analysis, it actually builds out [00:19:30] these dashboard like views for you as well that give you additional information as you're making those decisions. So we're going to see that with the different workflows that we have, whether that's AR, whether it gets more into payments, that's really the long term goal, is to have more of these workspace centered interactions within Sage Intacct.
Matt Lescault: Okay. I think we really need to address one thing because you said it. Copilot is a name, not Microsoft copilot. And I [00:20:00] think there's a confusion in the marketplace. And bear with me for a second as I ask this question, because it's going to be a little bit in a circle, but it'll all make sense. Um, why would a company like Sage invest in building their own AI within product when so many other organizations, just like Sage is offering the MCP? I know we'll get to that, where you can connect Claude into it, where so many other organizations are just sort of connecting LLMs into their product. [00:20:30] Why does that create a differentiation for Sage, and how do you think that really impacts the product and the utilization of that product. I know I went there, I told you to go in a circle, but it's all so related. I really want to get your your thought process there.
Taylor Brown: Yeah, no, I think that's actually a good question. Um, I feel like back to when I mentioned people kind of being almost a bottleneck in terms of AI adoption, people sometimes don't even understand how [00:21:00] AI can aid them in processes. It's almost like using a new muscle, right? So if I am using a muscle that I've never used before, I'm like, I don't even know what different scenarios and situations that I can accomplish with this new thing. And so I think with Sage actually building out more AI in the system and actually providing guidance for organizations on some of the best practices with using that AI, it creates a differentiation because it allows them to adopt these AI functionalities and [00:21:30] workflows in a way that is accessible to them. Sometimes when you're just connecting a outside model, People say. Okay, that's great, but they don't even understand the different ways you can use them in accounting processes. I think most of us use AI today for things like email editing and now looking up things. So almost supplementing a Google type functionality. So getting it more into work based workflows is going to need guidance, I think, from these organizations to provide to their customer base. [00:22:00]
Matt Lescault: I think that's really important to just, just for everybody that's listening to really take a step back. Because if you're just attaching LM, it doesn't know your data, doesn't know the product as well as when a product team like within Sage deploys this AI themselves. I just thought that that was important.
Doug Lewis: It makes sense. So if you haven't seen the finance intelligent agent in your intact environment yet, [00:22:30] I would say ask your administrator. If I'm not mistaken, it's rolling out now. So you might already be enabled. It might just need to be configured unlocked showing someone show you how to use it. But regardless, it's starting to roll out to the base. We'll see kind of what the adoption looks like, hopefully within the next couple of months here, I think we'll have a little bit more kind of data to go off of in that piece. So, um, a couple it's been teased a little bit. We mentioned MCP a few times. Matt, you threw, you threw clod out there. All right, I won't hold it against you. You jumped ahead. But that's that's okay. So [00:23:00] moves into the next part of the release notes here the AI gateway. Taylor what what is MCP? What even is that? What does that stand for? What does it mean? Why is it important to people. And shed shed a little bit of light please.
Taylor Brown: Yeah. So MCP, I think I would think of it as just like an integration point for AI models. So it allows you to connect clod with Sage Intacct or connect Microsoft copilot with Sage Intacct or a ChatGPT instance with Sage Intacct. And then you can use [00:23:30] the prompting in that tool to generate outputs. And that's going to be outputs within that platform, services within your Sage Intacct environment. Um, right now, what I've seen so far, and even for us, I think on the partner side, we're still learning about, I think what some of the capabilities are going to be from an MCP standpoint, but it's basically allowing you to use a trusted data source to ask questions and generate information so it can generate reporting. I know I saw an example [00:24:00] with the technical solutions consulting team at Sage, where they were actually able to use their cloud environment to create more of a customer relationship based dashboard view. And I think if your cloud based or your Microsoft Copilot environment is connected to your ERP and your CRM and your HR solution, that's to me where the real power in that type of MCP relationship lies, because then it has all of those data sources connecting [00:24:30] to one place.
Taylor Brown: And then within that cloud environment, you can generate reporting that really shows you everything kind of all together. Um, one way to, to look at it, there's different ways to kind of build out the connection points. You could say I have CRM flowing into Sage Intacct. I have HR s flowing into Sage Intacct. Sage Intacct is that main data layer. And then I do the MCP on top of Sage. And then I'm building dashboards based on the information that's there. You also have some people saying I have snowflake and that's connected to my homegrown system, [00:25:00] my ERP, my CRM, my HR s, my maybe donor management system. And then it's connected to snowflake and then it's pulling information from there. Um, so there's different ways to kind of, you know, go in and pull detail, but it's really about connection. And I would just almost say, if you're confused by it, it's just integrating with these outside querying type AI tools.
Matt Lescault: So Taylor, one of the things I was looking at in the release had to [00:25:30] do with some of the kind of licensure that is required to even connect to the MCP. What do you know about what people have to do in order to actually get access to and integrate with the MCP capabilities?
Taylor Brown: Yeah. So.
Doug Lewis: Well, Matt, are you talking more like on the on the actual technical security side of things or more on the, just the hoops to jump through to get whitelisted? You know what?
Matt Lescault: It's not necessarily whitelisted. I think it's the actual like licensing [00:26:00] that you have to have in product to actually connect through the MCP server. This is so brand new for us all, in my opinion, that there's just these questions about how to even do it. And I was just wondering if Taylor had any insight into that.
Taylor Brown: I actually don't I think we're still learning even over here. I it was interesting being at Sage because I was asking some of these questions as well about like the agent [00:26:30] marketplace. And I think Sage is still. I hate to say that working some of this out.
Doug Lewis: There pioneering new solutions as we speak. I think that's.
Taylor Brown: Correct. Yeah. I mean, I would say for anyone that wants to investigate it more, to look at the Sage developer site, and it's going to walk through the detail I think that's required for those MCP connections. I'd assume you probably do need a web developer license. They're probably just going to loop that in there. But I would reach out again to your administrator or whoever [00:27:00] owns your licensing, and then they should be able to get those answers for you.
Matt Lescault: I think, and don't quote me on this. I think there's like an MCP license, kind of like the web developer license. But I agree with you. Go to the web developer website and don't rely on us for this because it is changing. I mean, you brought up the, uh, you know, the, the agent marketplace, uh, which is brand new. Um, and I just had a little bit of, uh, of insight myself. I was talking to one of the individuals [00:27:30] that, uh, heads up, uh, that program and there's been, you know, uh, without sharing anything, I'm not supposed to, there's been, uh, almost 200 applications into that. Um, and Sage really is creating a wall or a gate and approving ones that are coming in. And I know that the number of ones that are being approved are very small as a percentage to the number of ones that are being applied for, which [00:28:00] I think is also a very good thing for the overall community to make sure that there's trust in what's being developed and there is continuity in what's being developed. So just on a on a side note there.
Taylor Brown: I guess I have a question for you on that. Did you get any insight, I guess, into how they're going to let the customer base access those agents? Because that's my question is, is it going to be something that you have to pay for credits as you use the agents or how those interactions are going to take place?
Matt Lescault: So at a high level, I know that you're going [00:28:30] to have kind of like a separate marketplace, like you do Sage in tax marketplace for ISVs. If and I don't have any insight like how they're going to structure the commercialization of pricing, but like you said, I think there's two options in the marketplace world, ISV World and Taylor, jump in, uh, any time here, uh, if I get anything wrong, but really there's a API cost that the marketplace partner consumes as part of their licensure [00:29:00] to, uh, and rev share to Sage. And so it wouldn't surprise me if that kind of rep model exists for this marketplace. Um, what I know is that all that is out has been approved is still in development. And so we, we won't probably see the marketplace open until next financial year is, is my guess. But I love for anybody at Sage to comment on this episode and [00:29:30] tell us where we got it wrong, because it isn't a very unknown area right now.
Doug Lewis: That's why the word unofficial does wonders.
Matt Lescault: Yes.
Taylor Brown: I mean, I think for me, that's the really exciting piece with Sage is the Agentic AI marketplace. Um, I think that's where everything is going. I was sitting at this presentation from this startup and they talked about that in the long run. Right now, we're more in a SaaS driven model world, but eventually it's going to be a genetic driven where we still need these SaaS tools that [00:30:00] the agents basically tell what to do. But people are going to decide what SaaS tools to get based on the agents that they have in place. And so I like that they're going to have an energetic marketplace. I think what's going to be interesting to me is the customers that aren't developing their own agents with Sage Intacct and how we're able to also share that type of detail with the other customer pools about just what's possible out there because obviously with the applications, you have a lot of the partners [00:30:30] building these agents or you have internal Sage building agents on the marketplace. But I actually think a lot of the power comes when a company is able to build out their own agent within their own ecosystem, because they might have unique needs that they need to take care of, which is a lot of the work we also do on the CLA side with our AI team. So.
Matt Lescault: So I think the number that I was given is that close to a third of the applications came from the customer base.
Taylor Brown: Oh very cool.
Matt Lescault: That's [00:31:00] awesome. Don't quote me on that. But I think that's the number that I was given unofficially and not for the unofficial. Yeah. Okay.
Taylor Brown: No, I love hearing that because to me that's they are in the weeds of things. So I, I really want to hear from customers what they're using on a day to day. I think it's going to spark others to be creative.
Matt Lescault: So, you know, a call out is if there's a customer listening that has sent an application in.
Taylor Brown: Yes.
Matt Lescault: Tell us about it and tell us the why?
Taylor Brown: Yes, please. [00:31:30] I'd love to hear about it.
Doug Lewis: And before we move on from AI, because we could we could talk about AI for 25 hours on here. All right, Taylor, since you're on the academic side and kind of the I call technical practice side actually doing it right. What is what is the worst AI use case you've ever seen? I mean, what have you seen? Just like a dumpster fire or a train wreck where you're like, oh my gosh, what were you thinking here?
Taylor Brown: Oh. I don't know. That's a that's a tough question because I [00:32:00] almost want to say just from like a people using AI on a regular basis, I do get concerned about people interacting with AI like a therapist because I think AI hallucinates and validates. And so I, I, I guess it's not a specific use case. It's it's past. Let me, let me take a step back. It's not just using it as a therapist. I think for me, the biggest concern and this goes back to, I think, our conversation starting about the students that are coming up in the younger generation is [00:32:30] we have them growing up in an AI world where all of us, we had to do things. I don't know how to write essays on paper. Right? Like we had to use Google like, and then AI came out, but we already had these strong foundation. But we're seeing a lot with students is students are using AI to write papers. Students are listening to Audible's students are doing their research through AI, but they don't have the foundation to check AI when it starts hallucinating. And so I think it's really, [00:33:00] for me, more so the concern with people not understanding how important, having a foundational understanding of the topics you're using AI for is to utilizing that in product, because then you can't verify if AI is doing a workflow correctly, if it's providing you with correct information, if a sentence structure looks right. And so I think that's what we're seeing now. And I don't know where we go from here because if people don't have that baseline, then it's [00:33:30] going to, to me, just extend out more and more misinformation and confusion.
Matt Lescault: So I think, I think what you're talking about is basic confirmation bias that you can get trapped into when it comes to the utilization of AI. Um.
Taylor Brown: Yes. Um.
Taylor Brown: The last thing I'll say to you, just from an accounting standpoint, AI is incorrect, like around accounting topics quite frequently, um, Claude is actually the best tool if [00:34:00] you're going to do any sort of AI or financial research and they actually think came out with their own like financial agent. Um, but I have to tell students that at the beginning of the semester, like, please do not run questions through ChatGPT. I have tested it and ran my exams through ChatGPT and it's gotten like 30, 40% of the questions incorrect. I think that also goes back to your question about why it's important for organizations like Sage to also build some of their own AI tools, because [00:34:30] it is accounting specific and accounting focused, and you have to keep in mind that a lot of these LLMs are being used for many different things. And so it goes back to being a mile long, but an inch deep versus having a tool that's a mile deep and then focused on a specific area. So when we think of stage two, eventually we're going to have AICPA information within the Sage copilot tool. It's going to be a better pool of accounting knowledge, especially as the language model [00:35:00] expands out and becomes more flexible, similar to a cloud in a copilot or ChatGPT.
Doug Lewis: And correct me if I'm wrong, but in the finance and accounting world, 30 to 40% incorrect makes a pretty substantial difference on the bottom line.
Taylor Brown: Yeah, it.
Doug Lewis: Yeah, certainly tough to pass an audit. Incorrect. Moving to the last portion here on the release notes here. Fixed assets. The audit [00:35:30] trail. Everyone loves talking about when things go wrong, right? Inevitably, they're going to go wrong. It is what it is, right. But Matt, where do you want to dive in on this one?
Matt Lescault: Well, I just I think it'd be really great, Taylor, for you to give us sort of a, an idea of why this is important from a fixed asset where we came from, because, I mean, I think Sage is original, fixed asset was a standalone product and now they've built this in product. And so I think a little bit of background to getting to where we are today and then talking about what the release note really tells us [00:36:00] would be pretty impactful for people.
Taylor Brown: Yeah.
Taylor Brown: So Fixed Assets has certainly been on a journey. Um, we started out with what was called the fixed assets tool, which is, was developed by the system integration team, which is still, you know, some organizations are using that tool. I think everyone is still enabled on it. So if you're using that functionality, that's fine. And what it was missing though was basically having fixed assets as a dimension. [00:36:30] So with the release of the Sage fixed Asset Management module, it now added in an additional dimension to your dimensional structure. When you get that Sage fixed assets tool. So that means I work with construction companies sometimes, let's say they have a piece of equipment that they're using to generate revenue, and they need to track time against that equipment. Now, on a time sheet, they can actually tag that asset on that timesheet entry. They can also tag [00:37:00] that asset on a vendor bill or on a order if they have work orders taking place. So now they're getting a fuller picture of the revenue that the assets generating any sort of maintenance that's taking place on the asset as well as the depreciation that's taking place on the asset itself, because they have that unique identifier that they can then tag insight and information to. Um.
Matt Lescault: Now that was all that was all there before this release, correct? [00:37:30]
Taylor Brown: Correct. So that was all their, um, part of the most recent release before this one was also the addition of construction and progress. So that was also why we saw organizations still using, say, sig fixed asset management. Um, because they needed that construction in progress. And now that is also within the fixed asset module. With this release, we got into things like reversing depreciation. So we have these. And I want to make sure I'm speaking about this correctly. [00:38:00] So we have these reverse depreciation entries. So users can now go in and reverse them if something is incorrect. So it creates a clear audit trail of information versus you having to go in and delete an entry. Um and then of course we have that audit trail Material functionality. Um, the first piece when we think about reversing entries, we talk about this a lot just with standard general ledger best practices. I'll have people ask me all the time, can I delete this? Can I delete this? And I'm always [00:38:30] hesitant, I think as an accountant to just delete, because then there's not a trail of what's taking place. I would much rather someone go in and do an adjustment or a reversal so we can see where the mistake was made, and then we can go back and see where it was updated and that change was done. It also helps to make sure that even if the error was made in a closed period, when that reversal takes place and the new period, we're able to really understand and reconcile that, especially when we look at our full fiscal year financial statements.
Taylor Brown: So I [00:39:00] think that's a big piece as well as we think about audit trail functionality, we have a trail now of these entries in the system to track any sort of mistakes that have been made. Um, they're also I think were some additional enhancements around splitting fixed assets. So it just made it a bit more efficient. As we think about splitting those assets from a process standpoint. Um, so I think it's just allowing us to have more detail from a trail standpoint as we look at our fixed asset management. Um, it's, [00:39:30] it's interesting. I just was working with an organization and they actually, from an audit standpoint, they had the most challenges around fixed assets and had received some feedback that they basically needed to get a tool that allowed them to have this type of detailed tracking to make it easier to investigate things when things take place, whether that's splits, whether that's mistakes, whether that's updating the cost of an asset, whether that's disposing of an asset or transferring it. It just makes it easier to have a better trail [00:40:00] of that detail.
Matt Lescault: Tell me if I'm if I'm thinking about this correctly. So like when we think about the split, because the newer product, not the SIG product, but the, uh, s fam, I think it is, um, you can, you can run fixed asset entry from the point of purchase. So from the Po into the bill market into the fixed asset, which means that if you have a $100,000 bill that has 15 assets on it, that split function allows you to do that in, [00:40:30] uh, in real time. And so that creates an efficiency. Is that, is that a fair assumption?
Taylor Brown: It is it's exactly. So it's like if I bought a 15 computers and I booked it as one line, and then I just say split it into 15 assets, it's able to go in and do that for me. So that certainly saves time versus me having one asset created and then going in and saying, okay, wait, now I have to split these into individual assets because they're all with different people at different locations.
Matt Lescault: So and as a partner, um, [00:41:00] and I don't know if other partners have this question, is there because there was that time period where you went to say it was construction in progress. And now with the changes or the upgrades that have happened here at SVM, do you see or do you ever find a need to position the Scig product still. And why would you, if you if you would.
Taylor Brown: So I haven't positioned it in a while to be transparent. I think for me, the indicator that we used to need to position it was construction in progress. So if someone [00:41:30] was doing a ton of, let's say a school, they're building four new schools in the next year, then we might need to go with Scig. Historically, with the updated construction and progress functionality. I am just going towards Sage fixed asset management because I do think that dimension piece is massive, and I actually think the construction progress flows smoother in the new Sage fixed asset management than it used to in the prior tool. Am [00:42:00] I actually tested it a month ago and it was very easy from the onset. Similar to when you create an asset from a vendor bill to say, this vendor bill expenditure actually is going towards a CIP asset and that's all you have to do is you basically just say, either it's a new IP or it's an already created one, and then you tag what the IP asset is, and then it just adds it to the record. And then it's as simple as the way you used to put assets in service. Now you just say capitalize this asset. It does [00:42:30] the entry for you. You can see all the detail very easily. So it felt just simpler to me to do that process and workflow. When I was demoing the old sig SIP functionality, I felt like I spent a lot of time in terms of prepping for those demos. It was almost nerve wracking.
Taylor Brown: To go forward you're like, okay, I need to follow like this exact workflow. And if something takes place like that's going to really be a big issue. Um, and I don't see that happening anymore.
Matt Lescault: Now [00:43:00] I know that current clients that have the SIG can stay on it and it's not end of life or anything else like that. But would you recommend that anybody consider migrating from the Scig product into S Pham at this point?
Taylor Brown: Yeah, I, I mean, I always think at the end of the day, I want to be on the most the tool that they're focusing the most energy on, right? Because that's where the enhancements and updates are taking place. [00:43:30] So I mean, if you're using a tool and your fixed assets are pretty straightforward, I mean, maybe stay on it, but if you feel like you have some time to make that migration Sage is clearly, I think, prioritizing spam more than they're prioritizing. I think the sig fixed assets. So I might and again, this is my personal opinion. I personally would eventually want to make that migration on the tool where we're seeing more enhancements take place if it's effective.
Doug Lewis: And so Taylor, we have, we [00:44:00] have listeners across the US, Canada, Australia, South Africa, France, a handful of other places that we love you all, by the way. Thank you. Is there anything in this release notes that you think is particularly meaningful for those organizations operating outside of the US? Well, Matt, maybe that's more up your alley since you're, you're you're. Mr.. What, what do we call you? Mr.. Half hemisphere because you weren't. Mr.. Worldwide. Mr.. Half hemisphere.
Matt Lescault: I think let's just go with I end up operating in multiple regions. So I have [00:44:30] a little bit more, uh, day to day interaction with how Sage Intacct is deployed from a regional basis. So I think, you know, what is most important to understand is that it's impossible for a global company to deploy all features immediately across all regions because of compliance and regulatory requirements within those regions. And [00:45:00] I think it's really important for any organization that either, uh, that operates internationally from a home base perspective, being buying intact, internationally, really focus on the feature releases of when it's coming to the region and what is coming to the region. And I think the biggest thing that I like to highlight is related to France and Germany and some of the e-invoicing capabilities that are being deployed in those regions, specifically for their [00:45:30] compliancy and how we expect that to, uh, to move in the wider, uh, Europe, uh, deployment of intact and how that is very important to businesses operating in those jurisdictions. Uh, so I just think it's really important for those companies that are purchasing outside of the US or really North America to, uh, to, to be very cognizant of what that, uh, what that looks like.
Doug Lewis: No, it makes sense. And if you are listening from outside the US, the regional release notes are broken up by country@intact.com [00:46:00] worth bookmarking. So you know exactly what's live in your own environment. Now. Taylor, I got to kind of move into this question here. We talked about a lot of a lot of different components in the release notes. You didn't get too technical, which is actually really fun for me. You are applicable. Maybe that's the teacher coming out in. You write how to actually use the information as opposed to here's the information. If someone's listening today using Sage Intacct and they can only do one thing differently after this conversation, what would you want it to be?
Taylor Brown: Read the release notes. [00:46:30]
Doug Lewis: Just straight up read them. Okay, now you're sounding like a teacher again.
Doug Lewis: Come on.
Doug Lewis: I read the textbook.
Taylor Brown: Um.
Taylor Brown: I mean, honestly, I would think just request to see some of the AI tools if you haven't seen them. I think that's the other piece when we think about AI is you need a more exploratory mindset. And I think from a SaaS standpoint, sometimes people say, well, we're going to look at what's out there every 2 or 3 years or every five years. And I think you're going to really miss out on [00:47:00] some operational efficiencies if you are not staying up to date with what is coming out, especially from an AI standpoint, because there's so many changes that take place, honestly, from a month to month basis. And so although Sage does the releases quarterly, especially if you're using in the long run things like the MCP and you're using clod, well, there's new functionality that's also coming out now in the large language model. So I would say the biggest piece is if you're not getting that curiosity to see [00:47:30] what AI is out there, even if you're not ready to use it, you need to at least ask or request a demo recording or read what's taking place. So you have an understanding of what those efficiencies look like, because especially when you get to the place with the financial intelligent agent where we're using more of the workspaces versus the dashboards, that's going to be, I think, a complete uplift in Sage Intacct, right? So many people start their day on those dashboard views. They're used to just classic reporting and then clicking into their workflows. Well, having [00:48:00] now this really interactive, engaging place to see your reporting, your KPIs, but also do your processes and not having the ability to use that and not even realizing it's out there. I think is going to be challenging in a crawl. Walk. Run is a lot easier as an organization than just making a big shift in a few years when everything is out.
Doug Lewis: Do you see a lot of people try that where it's like, hey, we're not going to start small, figure it out, learn how to use it and then get back. I mean, do you see a lot of people just jump head first into the pool?
Taylor Brown: I [00:48:30] think more people do a crawl, walk, run. I think some people mentally, sometimes when they jump into the pool and then there's so many different layers to that change in terms of what needs to take place. If they end up backing into realizing that they need to do a crawl, walk, run. Um, I'm sure Matt, you can also experience this like being in sales cycles when you have people that are like, and we're going to change five systems and we're going to build out all the integrations and we're going to do this and we're going to do this. And it's like they're just using QuickBooks right now [00:49:00] and nothing is connected. I'm like, oh my gosh, that's kind of a lot. Are we sure we want to do every single thing at once. And then they're like, And I'm hiring someone new and you're like, oh my gosh, like, let's take our time and maybe go step by step, like, what are the big wins? That's why we do to a lot of strategic mapping. So I always say, don't make a change unless you've mapped out what the next six months, year, three years, five years look like. And then the capacity of the team and [00:49:30] have an understanding of exactly what level of efficiency these changes are going to create. It helps you to prioritize those changes, I think, effectively.
Matt Lescault: I don't know what Taylor's talking about. I don't know who she's working with. I've never had that experience in my life. Kidding.
Doug Lewis: Oh. What are you going to do? Uh, Taylor, thanks for joining us. And for better or for worse, we have a tradition here. We end this thing with a terrible dad joke. Now, about half the time, I end up having to do it. And I just feel terrible after the work. If I'm not mistaken, [00:50:00] you might have one on deck for us.
Taylor Brown: I do.
Doug Lewis: So Yeah. I've never been more excited in my life. It is. It is with it.
Taylor Brown: All right. So it is a accounting joke. It's kind of accounting in AI. So if I asked AI to tell me an accounting joke, and here's what it said to me, it gave me a 14 page report. Three assumptions, two disclaimers and a confidence score of 72%.
Doug Lewis: Well, [00:50:30] if the goal was to tell a terrible joke. Taylor, you nailed it. I couldn't have done it better myself. And it was great at the same time. And I absolutely love that. That is a phenomenal high note. I'll call it a high note to end this one on. But Taylor, thank you so much for joining us and we hope to have you back here in the future.
Taylor Brown: Thank you.
