AI Meets ERP: What Acumatica Partners Are Seeing in the Field

[00:00:00] Andrea Hays: We're in an age where things like intelligence and data are ubiquitous, right? They're everywhere. And so what becomes scarce is trust.

[00:00:07] Rob Kirkey: If an AI agent is the one who created a transaction in- incorrectly, you're not gonna get a get out of jail free card by telling the auditor, "Hey, it's AI's fault. Wasn't my fault."

[00:00:17] Armando Huerto: Now it's not about what can't AI do, it's what it shouldn't do.

[00:00:30] Stephanie McColly: Welcome to the Acumatica ERP podcast, where we explore practical innovation for growing businesses. I am Stephanie McColly, the chief marketing officer and EOS integrator at Crestwood Associates. My team is a longtime business operations advisor and Acumatica channel partner, helping businesses implement and get more out of Acumatica's cloud solution.

Through this work, we have firsthand view into the opportunities and also the challenges of growing businesses that are trying to navigate every day. Now, AI is on every business leader's mind right now, but the real question is whether it actually delivers once all of the hype settles around it. So in this episode, we're gonna dig into three big areas, the first one being how AI is showing up in businesses today, where it's creating genuine value, and what organizations should weigh as they figure out their own path forward.

And so to unpack it, I've brought an entire band here with me today of three Acumatica partners, each skilled in business advisory, but all with different vantage points. So give us a wave here as I introduce you. Rob Kirkey, chief revenue officer of 3Value. Thanks, Rob. Andrea Hayes, principal for SVA Consulting.

And then Armando Huerto, chief executive officer of i-Tech Support. Thanks, guys. All right. So between them, they bring decades of experience helping businesses adopt ERP across industries as varied as defense, regulated life sciences, and complex manufacturing. We've seen AI land in a lot of different rooms in our roles, and today we're gonna share what's really working and maybe a little bit of what isn't.

And so welcome, team. Thanks for being here. I'm going to get us rolling into the first question. I've got a series of questions I really want to dig into today, so everybody hold onto your seats 'cause we're in for a ride. Before we get started, let's do some quick laps around the table here. I want to talk about three key things about each of you, and Rob, I'm gonna start with you.

The first one is tell us about your firm, how long you've been an Acumatica partner, your primary industry focus And then your biggest takeaway with how AI may be impacting your customers right now.

[00:02:29] Rob Kirkey: Okay, thanks Stephanie.

[00:02:31] Stephanie McColly: Mm-hmm.

[00:02:31] Rob Kirkey: 3Value's been an Acumatica partner since our beginning when we started in March of 2020 with one employee, who was our CEO, Nick Knight.

No install base and just in time for COVID. Since that time, we've grown to over 150 clients, largely made up of manufacturers and construction companies, as well as organizations in the highly regulated firearms, defense contracting, uh, nonprofit sectors. Our clients have been impacted by AI in the same way that AI is impacting 3Value.

You know, in January, most of them couldn't spell AI. Now, almost all of our clients are aware that they need to know how can AI assist them, how can AI be safely deployed in their organization, and how to protect themselves from malicious AI-based attacks on their mission critical information and systems.

[00:03:19] Stephanie McColly: Yeah, that's a lot. That's a lot. And I know that, like, in the regulated side of things, Andrea, you're also kind of playing in that space. So why don't you give us a quick rundown on your firm, how long you've been a partner, that industry focus, and your biggest takeaways for AI?

[00:03:31] Andrea Hays: Yeah. So Andrew Hayes, the principal here at SVA Consulting.

So we've been in the ERP world for about 30 years, so since before my time. Before we came on with Acumatica, which was in 2010, we were implementing, you know, Microsoft SL and, and GP, which are now, uh, legacy systems. So at this point, we, we really focus in, in, on two buckets, right? So we have a lot of product-based businesses like manufacturers, dis- distributors, and then we also have a lot of project-based businesses, too, like professional services firms, and one of our specialties is in that FDA, that regulated space.

But really, what our customers have in common is they have operational complexity. They've typically outgrown their legacy systems, and they need a, uh, a, a way to connect into their business a little bit better. So in terms of Stephanie's last question, right, the biggest takeaway for our customers in AI right now, really seeing a lot of those, those trends in terms of, you know, efficiency, so starting to eliminate or at least reduce the need for a lot of that repetitive task, and in the way that information is consumed, right?

So especially with those old legacy systems and moving to Acumatica, it's a lot easier to get access to the data that you need to make decisions at the speed of business. And I think AI is taking that to the next level because now I can interact with the data, I can ask questions of it, and I can glean a lot of insights a lot faster than I could have before.

[00:04:44] Stephanie McColly: Yeah. Those are great. We're gonna dig into so much there. All right. So Armando, last but not least, of course. So give us the rundown of your firm and, and all the details.

[00:04:52] Armando Huerto: Yeah. Thank you, Stephanie. And, uh, so i-Tech Support is a managed services Acumatica ERP provider, and we recently launched an AI-as-a-service offering specifically to answer the beck and call of what we see a big opportunity for clients in our space.

Um, our company's about 75 people strong. We have roughly 300 clients across the US, Canada, and the English-speaking Caribbean. We've been an Acumatica practice since 2011, so we've been in business for 15 years, and our three service lines really answer the, the question of how can we be a technology partner to our clients?

But more importantly now with the addition of AI is really how can we be an intelligence partner to clients? And I think that that's the biggest impact that artificial intelligence makes, is that intelligence really comes from the human side and the artificial side and the infrastructure layer alongside of it.

So I think most partners that- Can get that right, that can be a consultative really clarity or give the ability for clients to actually navigate the intricacies of it and deliver value are going to be the ones that can provide the impact. So to me, leading with value and deploying AI in a governed way is, is really the way that we're trying to approach it.

Our industry focus is for, for us, is primarily manufacture and distribution, which are core to Acumatica. We also have a, uh, a big focus on construction and, uh, service trades, so more specifically commercial landscapers, which is a very interesting way that we're deploying AI because there are different systems and layers of software that a lot of these, uh, a lot of these companies utilize.

That is a great opportunity to, to find ways to better integrate systems. And as Andrea and Rob said, govern it but also ensure that the data is well structured so that you can actually interact with it.

[00:06:39] Stephanie McColly: Yeah. The governance piece is one I know is very popular, especially with the CFO crowd, who is sort of being in charge of, uh, or being charged with the objective to go out and figure out what AI looks like for the company, right?

And so I know that security, safety, governance, all of these things are, are critical. And it's interesting, you know, on our side of things at Crestwood, we joined the Shield Technology Partners portfolio, and they are an AI operations platform that's actually backed by Thrive Holdings. It's the firm OpenAI has taken an ownership stake in.

And so currently, on our side of things, we're working with Shield and their team of engineers, and we're kind of rethinking, like how does business get done using the ERP as a system of record for AI tools? And so it's almost like going back to startup mode, sort of reframing your thinking and rethinking your business from the ground up, which sounds exhausting, but also incredibly interesting because you get to sort of revisit the business structure.

And so I'm curious, you know, Armando, given that the, the connection between AI and ERP, you know, how are you actually seeing those two impact you, um, with your business, but then also the primary industries that you're focused in with maybe manufacturing and distribution?

[00:07:44] Armando Huerto: I think there's, there's kind of two things that I, that I see is that folks are expecting AI to provide some sort of easy button into whatever it is that they see and then identify, but that also becomes a lot of noise.

And that happens to us, frankly, right? We, we want to automate, we want to reduce manual work, we want to elevate our people into doing more value added services. But I think what is really impactful across all different industries is that we need to make sure that we do this in a controlled manner and in a way that you have- A named value or really a cost associated with what you're trying to automate or implement AI to, and also en- ensure that you have ownership around it because I think of AI still, and, and I continue to do that, as a, a really a, as a support system or really an analyst and not just a fully autonomous individual or entity that can do everything for you without any oversight.

And I think that part of that, and I think we touched about that earlier, like bringing it back to full circle, the folks that we've seen go through a readiness assessment that we run that have really leveraged a full-blown data system like Acumatica, and that is their system of record, can turn into the system of intelligence a lot easier than those that maybe aren't necessarily leveraging all of the features or all of the modules that they could in their business.

So I think you guys probably see this a lot where we do phased implementations or we, we, we maybe run WMS somewhere else, and the more disconnected the systems are or the more the data is scattered across different places, the more difficult it is to actually create a process that is gonna be actually, uh, automated, right?

That is gonna enable you to, to, to automate. And so in summary, I think that the impact that I see today is that the underlying assumption is that AI is going to be an enabler of, uh, of value, but I also have not seen a very good us proven way to deploy it. Mm. And that's where I think us as partners need to work alongside our clients to do that in a way that feels like it's custom to them, but really specific to something that they need to chip away at, and not try to take a bite of the whole apple in one fell swoop.

[00:09:57] Stephanie McColly: It's kind of interesting that you say it that way too, because I'm reminded when it comes to somebody in manufacturing or distribution, whether you're talking to an operations lead or maybe a financial lead, they know their business. They know the data that surrounds their business. They know their business processes.

They don't know AI usually, right? And maybe they're gonna kind of skill up on it, but wow, what an importance of a, a partner like an ERP partner like us to come in and say, "You don't have to have it figured out. You don't know what you don't know. We have resources that we can tap and we can help with this.

And by the way, we've got your back with the foundation of everything, which is ERP," right? So it's interesting. It kind of is like we become the easy button for them.

[00:10:35] Armando Huerto: No, and I think it, it's, it's more about that partnership of they have to... You said it yourself, they know their business, but they're also trying to run the business.

[00:10:42] Stephanie McColly: Yeah.

[00:10:42] Armando Huerto: And they're getting pressure, whether it's innate or external or even within themselves, that it's almost like, uh, like they're getting left behind. I, I say keep up- Mm-hmm ... with the Joneses, right? It's like, well, if everybody, everybody must be doing AI. The reality of it is, is a lot of people are using AI, but it's largely ungoverned and it's largely more piloted and kind of ad hoc.

And I think that the opportunity for a lot of, for a lot of companies is to take a step back and think of it as a, not just like another project, but it's, it is a change management project and an opportunity to set the foundation so that whatever it is that they're working on, it's actually gonna be long-lasting and repeatable.

And that's where I think we can bring a lot of that clarity because we can do it across different industries and bring the best practices. It's very similar to ERP implementation in a lot of ways. It's complex, but people assume it's just, "Oh, okay, let me just get ChatGPT, and my life would be solved."

It's a little more nuanced than that, and that's where I think that, you know, clients being open-minded about that and also being willing to share what they wanna share and, and having a structure is gonna be really important.

[00:11:44] Stephanie McColly: Yeah, super insightful. And you mentioned the governance and, and some of those pieces.

And so Rob, I actually wanna come to you next because this really kinda dovetails into a question I have for you. You know, you work in a pretty demanding, highly regulated world with your clients. How do regulations extend to the adoption of AI? What does that look like in your world?

[00:12:04] Rob Kirkey: Yeah. I'll, I'll take the, uh, most extreme example, which is the 700,000 defense contractors that are in the United States.

You know, they work with contract unclassified information, federal unclassified information, and because of that, Department of War has come out with a set of regulations under the acronym CMMC 2.0. It's a cybersecurity set of requirements, and how they relate to AI is that for a defense contractor, any AI component that they use in their, uh, AI technology, including a large language, uh, model, an LLM, or a vector database that processes that information, um, that would be available in their Acumatica Cloud ERP or any other mission-critical applications.

They need to meet the same 110, yep, one one zero, security controls from the National Institute of Standards and Technology that apply to the organization's deployment of Acumatica Cloud ERP, for example. Um, on their data's, uh, center side, there's, uh, security requirements under an umbrella called FedRAMP, and just like their Acumatica Cloud ERP deployment has to be compliant with that, so does all of the AI components that they use.

They cannot be having- AI models, uh, outside of that secured, uh, compliant environment. They cannot be calling them through API calls from applications within the environment. And then if a, uh, contractor works with components, finished goods, or even has information that's considered defense articles, then the AI technology that they work with are gonna be subjected to a compliance requirements with the, uh, initials ITAR, ITAR, which stands for International Traffic of Arms Re- uh, Regulations.

ITAR requires that AI technology that the client or customer is using, including any AI subprocessors that are being used, uh, in that stack, and the data that they access has to be located on US soil, has to be accessible only by United States citizens or green card holders. And the last set of initials I'll throw at you is DCAA.

This is a list of requirements that stipulate how the exe- execution of work is recorded under federal contracts. DCAA auditors must be able to trace a transaction from the point of creation all the way to the GL. And if an AI agent is the one who created a transaction in- incorrectly, you're not gonna get a get out of jail free card by telling the auditor, "Hey, it's the au- AI's fault, wasn't my fault."

As a result of an AI agent's misallocating labor incorrectly, a contractor could lose their contract, they could incur fines, there could even be jail time in- involved. So it's a very serious, uh, set of things that the people in the defense contracting industry have to consider when they're evaluating how they're going to a- adopt AI.

[00:14:59] Stephanie McColly: Yeah. It, it's interesting to me because w- what you're explaining is really kinda two sides of the spectrum. One side is highly secure, very governed, very minimal approach maybe that can be taken to some extent. But then also you've got the other side of the world that really is like, "Hey, let's build all sorts of agentic models.

Let's maybe set up some AI automation." And it sounds beautiful, but, you know, what is real really kinda comes down to the basics of the business needs, what their priorities are, and really what they commit to their clients or their end users. And so I can see how there's just such a varied approach. And so, you know, speaking of that at kind of the ba- the base level here, you know, Andrea, as we think about how we advise customers, like, when your customer asks where should they start with AI, what is it that you're telling them to consider?

Because we just heard one sort of story there, you know, that kind of tells a storyline, but what are you hearing and what is your advice?

[00:15:52] Andrea Hays: Yeah. I think where to start is, you know, the foundation of the business first has to be, you know, solid, right? So what does that mean? It means a really good system of record and, and that's why we like working with Acumatica.

And what it allows us to do is, you know, it has these open APIs that we can extend it very easily, we can connect it to other systems. And what we're seeing also is this kind of breakdown between the walls of SVA has a, a traditional BI team that needs to do EDWs and data lakes, and we're seeing kind of a breakdown of that distinction between ERP, AI, and BI because it's, it's one outcome or one solution that we're looking for for the customer, right?

So I think what this, what this lets us do is build things faster, build things better, and then achieve a, a faster and better ROI. So I mean, that's the first thing, right? If you're on an old system like SL or GP, or if you're trying to grow really quickly and you're in QuickBooks, right, you don't have that solid f- You know, foundation yet to add onto.

So that's really the first step. The next one is we also do AI assessments as well with our, our customers, and it's in a, kind of a nice scripted survey set of interviews to understand, well, where's the friction in your work today? And that lets us see, okay, well, what i- what's the areas in terms of the, the bottom line that we could start to improve on?

And then also we try to understand what's the top line as well, right? So what are the things that you're not doing today that maybe you could, or you perhaps you wish you were doing that could, could drive growth and drive revenue? So we really try to t- look at it from, from both of those lens. And in terms of where to start, sometimes that assessment can be helpful, but a really good use case for AI specifically, uh, to look at is when the risk is low, right?

It has to be use case where mistakes are cheap. Start with a question, not an experiment. So we're, we're, we're finding that, like, if you just explore without a question or an outcome or a vision in mind, you don't have, like, the, the gates around it, it's really hard to get meaningful action and learnings from that.

And really start small, right? Start with something simple, and then something that's measurable as well, right? So how do I know it, it worked? How do I know it didn't work? So really, I think it's the small projects that help us kinda move and, and see value quickly. I mean, that said, it's, it's kinda interesting that when we do a lot of these assessments, what we find is, you know, what shakes out is, yes, you know, there's some, some need for AI here, but actually a lot of it might just be deterministic logic that is just, you know, not a business event in Acumatica or automation or perhaps an integration.

So it's interesting 'cause it spawns a lot of discussion of how to optimize a business and reduce friction. Sometimes AI is the right tool set for that. Sometimes it's not.

[00:18:12] Stephanie McColly: Hmm. You know, you make a good point too because, you know, I'm curious with, like, the AI assessment, is that something you guys have done internally as well and you're kinda looking at your own internal processes?

[00:18:23] Andrea Hays: Yeah, we do. And it, I mean, it's, it's, it's interesting and it's kind of a fun time I think 'cause we're really looking at, at our customers and helping them, you know, scale, and then also looking at ours and saying, wow, like we need to revamp this, this component of the business. So yeah, it's- It's making us examine, uh, a lot of our systems.

Well, yeah,

[00:18:39] Stephanie McColly: we joke around that we're the cobbler's kids sometimes, where, like, everybody else has really nice shoes, but, um, us at Crestwood, maybe they're not as nice. Uh, so we've really kind of taken an investment into this one, and we created what we call the AI Strike Force at Crestwood, and it's a team of individuals that are just really inclined to just go figure things out and get creative, and it's been a lot of fun to see it.

One of the sort of use cases that we've put together is on the marketing side of things, because it's where I saw the greatest opportunity for us internally. And, you know, we built out a system of seven, it'll be nine agents soon, that are agentically working through our marketing engine and delivering some real value right now, whether it comes through the web traffic as the ROI and the conversion on the web, or really productive email campaigns.

We're, we're seeing some interesting things there. So, you know, for each of you, I, I want to start here, and Armando, I'll start with you. You know, can you share a use case maybe where AI is really delivering real measurable value right now? You can give it a number, and it, it could be a number of whether it's saving time, it's improving accuracy, or maybe it's just helping to create better decisions.

Could you give us, like, a, a rundown of one?

[00:19:45] Armando Huerto: The obvious answer has been, and I think Acumatica had solved for them pretty early on with, with AP and AR. I, I think that invoice capturing, automatic, automatic, uh, payments and things like that are, are, are really, really big. But what's interesting about that is despite having those tools, a lot of times they've been created with kind of the broad application in mind, right?

There are a lot of instances in which clients have specific processes or maybe even the date the input or the OCR optical, uh, character recognition that AI leverages to be able to manage AP and A- AR automation is not as robust or you have to train it for specific things. And so they're really nice out-of-the-box solutions out there, but there are some

in some cases that it's nuanced. So one example that we're really proud of is that we, we had a client that was been with us with Acumatica for over 10 years and was still manually keying in some sales orders and purchase orders that were in- e- eventually leading to actually creating sales orders and purchase orders for materials.

So they had one individual processing 25 to 100 of these a month, despite the fact that they had AP document recognition and some of that. Mm. And the reason being is because in practice you get these documents that are handwritten, that are, you know, maybe s- you get a little chicken scratch or something on that, and it all sounds really good on paper, but if one breaks, then now you're gonna have to insert a human.

So one of the things that we did is we created a actual ... We have a workflow automation platform called Alloy, that what it does is it uses AI to basically write a workflow and code so that it's more deterministic rather than event based, right? So it doesn't treat every s- it treats every single kind of issue depending on what it finds.

So what it's doing is it's filtering the cases where it's just an easy, "Okay, I'm just gonna, I'm just gonna input this in the system, no problem," and it just flags the ones that it doesn't know, but it verifies with a human. And that has been a really big awakening to show that it doesn't have to be an all or nothing, right?

It doesn't have to be a black and white or this solution works, is that now we can use agentic AI to create deterministic workflows that are saving this company. You know, f- if you're doing 50 of these a week, right? And we did a little bit of the math, it turns out to be no, no less than about $5,000 a m- a week of just time savings that that individual is getting back.

And all that we had to do was really work with them to work through the problem. And so One of the things I, again, just to summarize, is I think that now AI opens the door for us as partners to really not just be constrained by what we have out in the market or solutions that are, quote-unquote, out of the box, where now it's not about what can't AI do, it's what it should it do.

And that is a really good example of where we kind of have best of both worlds, and we're actually putting some real value into it.

[00:22:47] Stephanie McColly: Yeah. There's, there's so much opportunity, especially as you think about meeting people in the flow of work with solutions that make sense for them. We just launched our sales order agent, and it went from taking 10-minute orders to get created down to seconds now.

And it's a, you know, an Acumatica add-on. And it's one of the things that as we started building it out, we realized there's an opportunity to solve real problems. And actually, if you start to look at the industry and you get really niche into specific industry pain points, you can solve very critical things for, let's call it, process manufacturing, right?

And that's where we kind of specialize in. And there's really intricate solutions there that AI can solve if you really, really know the industry. And so I, I think that's a lot of fun for people, and it kind of shows up in different ways, too, depending on what you are processing, whether it's chemicals, and maybe it's regulated.

Perhaps it's, you know, um, powders. It's, it's all over the place, but it's super fascinating to me. Rob, you know, on your side of things, when you think about a use case for AI, what's really kind of delivering real value, like measurable? Give us some numbers.

[00:23:50] Rob Kirkey: I think it's, uh, you know, kind of building on what you just said.

You know, we also are building products that complement Acumatica Cloud ERP, and one of those use cases is a smart estimator. For our m- manufacturers, uh, their customers will often submit a, a drawing and say, "Give me an estimate." That simple question can take hours and, in some cases, days to convert that, that drawing into a bill of materials and an estimate within Acumatica Cloud ERP.

The, uh, product that we developed using AI now turns that multi-hour, multi-day process into a couple minutes. Um, so real substantial value

[00:24:27] Stephanie McColly: How do you have the conversation with maybe on the client side to help them realize the value of that solution over time? What does that look like for you?

[00:24:35] Rob Kirkey: Pretty easy.

You know, we ask them to, uh, share a drawing that one of their customers shared with them. We bring it in, and that's where, where I got that date, 'cause I- o- one of the estimators said, "That would've took me two days to do." We were able to show it to him, deliver within minutes.

[00:24:51] Armando Huerto: You make a really good point.

Sorry to interrupt, but because of AI, now we can really show something so much quicker. Things that used to take maybe, oh, should we go to this vendor or should we go to this solution, now we can kind of... It's almost like a little scienc- science experiment at times, but we can turn something around in a couple of hours that is pretty darn close to not just a proof, proof of concept, but almost the solution.

And I think that that's what- Right ... clients are excited about and they're expecting. But then we're also learning as we go on what is the art of the possible, right? Oh my gosh, like I didn't think they could do it this quickly. And so I think that that's what floodgates get really fun, and that's where I think that's a really important part, is our team members are really excited about it.

'Cause now they can put what I c- like to say, points on the board really quickly. There's like this level of satisfaction, almost euphoria, of like the endorphins that, oh my gosh, like I got to do this very quickly and very, and, and in, in a very interesting way. So that's also, I think it's exciting for our team specifically, and also for our clients.

[00:25:50] Stephanie McColly: Yeah, I have a need for speed. I love the speed of solutions that I can build because, um, I, I like to have lots of spinning plates, but I also have to push them through the finish, right? And I think a lot of times in businesses, you'll come up with an idea, you're feeling creative, you get together as a team, but then it kind of fizzles out a little bit later, and things don't actually get fully executed in some circumstances.

And so, you know, Armando, on that note, what do you think is really kind of separating businesses that get real results from AI versus those that kind of stall out after they get excited and have a little brainstorming session?

[00:26:20] Armando Huerto: I wanna say it again for the people in the back Start with a process, not the technology.

A lot of times people are coming to us saying, "I want this module," or, "I want ChatGPT," or, you know, or, "We bought licenses for this and now we expect XYZ," but they rarely have thought about establishing a baseline, really understanding the process, or even making the process eligible for AI to be used, and the process includes the data, the people, and the technology associated with it, right?

I think it's, it-- You know, if you cannot establish a baseline, I think Andrea said that well earlier, where you cannot state what the cycle time for something is today, how many, how-- the error rate, you really can't measure it. And so if you can't measure it, you don't really know what you have, and you don't even know if it's worth your time.

Even if it takes a couple of hours to automate something or to create an AI agent, is that a better output or a better ROI than doing something that's gonna be the same amount of time, but it's going to be significant and a lot more, more ROI? And, you know, I, I think that scope, narrow scope is gonna be something that be-- And I mentioned this earlier, like our team is excited, clients are excited, users are excited.

But start with one thing. Run it for 90 days. Put it into production. It's gonna fail. There's gonna be some trials and tribulations, but that forward positive momentum that you're gonna gain from that one process is gonna give you not only the confidence, but the ability to scale that a little bit quicker.

A lot of people try to Start kind of big instead of small. And then I think, you know, Rob, you mentioned that obviously you're deep into the governance stuff, but even SMBs and all businesses should, should ensure that they train their folks, that they have an acceptable use policy, that they know where the data is going, and that there's a human in the loop.

You know, the winners that treat AI as an operational change with a technology component and not a technology purchase with an operating side effect are gonna be the ones that win, and

[00:28:14] Rob Kirkey: I

[00:28:14] Armando Huerto: firmly

[00:28:15] Rob Kirkey: believe that.

[00:28:15] Stephanie McColly: Yeah. Uh, you make several solid points there too, because you really have this spectrum of employees, some who have really skilled up on it very quickly, others who maybe are g- a little further behind.

And it kind of comes to this place where, like, you can look at your business processes and you can say, "Where should we go from here?" And some employees just don't know what to do next, right? And they have to have the space to go test things out, play around with it, and try to figure it out. One of the things I recently did with our accounting team is I had them put together an AI wish list.

I wasn't sure how much they knew about AI, and I wasn't sure how much they had been dabbling in it, but I know that they know their pain points really well in their day-to-day processes. And so I said, "If you could wave a magic AI wand in the finance operations at Crestwood, what would your wish list be?"

And they sent it to me. And so now that's our project list that we're going through. And what we're realizing is there's some basic fundamental things, because we use Acumatica internally. We, uh, we just need to tweak the system. And then there's a few other things where we need to create maybe a workflow in Acumatica, but then there are true AI solutions as well that we can also be building with.

And so it was eye-opening to me to see, don't ask what you want AI to do, but go back to your process and ask what do you want your process to be like, what pain points are you experiencing, and then let's go solve for how we get the best fix for that. Rob, I see you shaking your head. You wanna add anything to that?

[00:29:36] Rob Kirkey: Um, w- we're doing a similar thing in that, um, you know, we have a, a company-wide initiative that, that has the acronym BRAIN. And, um, we started very small, but now we've gotten to the point where we're developing a brain where we're evaluating all of our organization's, uh, business processes and res- redesigning them to take advantage of AI where it makes sense and, and can add benefit.

[00:29:59] Stephanie McColly: Mm. Yeah. Beautiful.

[00:30:00] Armando Huerto: Guys love your acronyms. It took me a minute, but, but brain is clever. I, I've lost track. Brain is clever. I'm, I might, because it has A, A and I in it, I might have to steal that one, Rob. I'm sorry.

[00:30:10] Stephanie McColly: It's already a good one. I've lost track at this point. I like it. But I think the whole alphabet's in there probably.

Um, Andrea, you know, on, on common misconceptions, you know, I spoke about that with the team and, and sh- really trying to understand AI. You know, what kind of common misconceptions do you run into about what AI can and can't do for the business?

[00:30:28] Andrea Hays: Yeah. So I think sometimes it's, it's easy to just think of it as a cost-cutting tool, and sometimes that's, you know, the...

It, it can be, right? You can optimize a lot of, a lot of process, but, you know, it's, it's also really tough on employee morale to say, you know, "We're not gonna hire new people, and we're, you know, gonna eliminate your role over time," right? So I think it has a lot more value beyond that. And so when we start to think about what, like, what the misconceptions are, right, it's, it's that it can take over a lot of components of the business.

I think it can, but we're in an age where, uh, things like intelligence and data are ubiquitous, right? They're everywhere. So then, then what becomes scarce, right? And so what becomes scarce is, you know, trust. It's judgment. It's the bil- ability to take the data that comes out of, you know, Claude or ChatGPT and able to talk to it in discernment.

So, like, I use those tools every single day, and I, um, I constantly argue and question with them, right? To say, "Hey, is that, is that accurate?" Or, "Cite your source," right? And so, like, they can hallucinate. They can be wrong a lot. And so I think it's easy to see these tools as, wow, they have access to so much data.

They're so much smarter than me. But in reality, I think there's y- we're, I think we're missing or perhaps we're staking the value that the human has in the equation. I think that has to be kind of front and center, especially as we move forward, like, in the, in the age of AI. Well,

[00:31:40] Armando Huerto: those are a lot of- I don't know if you guys saw that EY now gives humanity type of bonuses, so when things like empathy or adaptability, because there's been so much AI slop that- Yeah

they're trying to get people to kind of go back to what made them human, and like, actually... So I, I found that fascinating. This was like, I saw this this morning, so what resonated is exactly what you said, is that access to data and, and processes is almost, like, ubiquitous, but at the same time, you need a human in the loop to be able to have the right sound judgment.

Bringing that back to what Stephanie was saying is industry knowledge, right? I think that's what our clients, anybody that's listening to this podcast, is desperate for, is like, "Can somebody come in here and tell me how can I gain a competitive advantage?" And you can't just do that with, with data. It's going to be a function of it, but it's how you interpret the data and make the decisions with judgment and with care that I think is kind of differentiate how folks add value, but also how business leaders make decisions.

So I totally agree with you with that on Andrea, and I found it so interesting that EY was doing that.

[00:32:41] Andrea Hays: Yeah, I think a lot of it too is how, you know, what, what makes your business unique? What's your competitive advantage? And I think if, if used the right way, AI can really expand that very well. But if, if not used the right way, right, if I just take its advice and move forward, you know, without discernment, that's, that's where we get into trouble.

[00:32:57] Stephanie McColly: Yeah. You know, with Armando, when you think about where AI's heading, what do you think you're most excited about? But also, what are you kind of watching out for that might maybe concern you?

[00:33:06] Armando Huerto: Yeah. I mentioned this to my team internally because as, as we were rolling out our AI strategy, I, I, it took a lot of time, but the thing that really captured me was that really for the first time ever is that you don't need an enterprise level budget to, for a 200-person manufacturer to deploy something that is incredibly meaningful.

We're genuinely leveling the playing field, and I would argue that it's easier to deploy AI in a smaller organization than a really bureaucratic one. There's countless instances where big corporations like Starbucks have, you know, deployed AI and completely failed because they're just very large and very clunky.

And that's what I love about the space that we're in. The other piece, going back to ERP, is that ERP's becoming increasingly conversational. Like, what I mean by that is that the expectation should be that you should be able to query your system one way or another and find the data that you need, and perhaps even think with it, alongside of it.

And I think that that's That's going to augment what we can do with ERP in a way that I don't think any of us really even comprehend today. And I think selfishly, like it makes the partner ch- channel matter more, not less. I think folks that someone that has connected their data, they're gonna need someone to secure it, to help them manage it, but also to help kind of the put points on the board that we've all been talking about.

So that's what gets me the most excited. I think what worries me is what a lot of people fear is, is the security and the governance, right? Uh, we're a managed services provider. We, we constantly see folks installing Claude on their machines, and we tell them, "Hey, this is how much AI is being used in y- in your environment."

And, and leaders are surprised. And that leads to things like shadow AI, right? Where if you don't have a policy, if you don't have training, that that becomes a risk rather than an asset. I think today AI is not thought of as a cost. I think in the future, and probably in the near future, tokens is gonna be like our power bill.

It's just going to be something that sits in your OPEX that you are consciously investing in. It can also lend itself to a dependency on one platform or on one vendor. And I know that when we met at the Acumatica Summit, I wasn't even a Claude user and Claude wasn't even mainstream, and here we are a few months later and it's all the rage.

And so I think sometimes I think that that's, that could be a, a, a risk for people to manage, to not dive into one ecosystem until we kind of find the winners and losers, and to keep themselves relatively flexible to that. And the last thing is, you know, it's not that like AI keeps me up at night because it's making as many bad-- Like what keeps me up is, is that AI is not making a-- implementing AI is not a bad decision.

It's that a company that cannot explain why they implemented AI and why the decision was made. Those are the ones that can get, it can get really tricky and it, you can go down a rabbit hole You know, it, it, it can become a- an added risk rather than an asset.

[00:35:59] Stephanie McColly: Yeah, you know, you spoke about the usability and the importance of being able to easily access your data.

We know that financial visibility is one of the biggest pain points for financial professionals, just to get the data out of their system. They know it lives there, but they need it in a way they can understand it. My entire career working in, like, building systems, I call it one-click solutions. You can create now a one-click solution where you can just visually see everything you need in just one click.

And man, it just really streamlines the entire experience for you at work when you can have things that are frictionless and give you the data that you need when you need it. And I really think that's the importance of the ERP being the foundation with very structured data, clean data. You know it's being maintained, the hygiene is there, and you're able to...

In, like, Acumatica's instance, you're able to really depend on the data that's in the system and not have to fight between teams to come up with what does that number mean, right? So internally for you, Rob, how is AI changing the way that you run your firm from maybe, like, your teams to even the services that you're hiring?

[00:37:00] Rob Kirkey: Within the brain, we've got it from one end to the other where our BDR has, uh, AI agents helping him identify which folks are the ones that he should be interacting with. Our sales team, um, as they're having their initial discussions with our prospective clients, agents are going out to that prospect's website, pulling down their logo, pulling down finished products Product descriptions starting to populate a sample database so that when the seller gets to asking the client, "Would you like to have a look at Acumatica?"

It's all populated with their data, so they're gonna be able to envision exactly what Acumatica would look like in their organization. When we move from the purchase to the kickoff, we have our consultants using AI to search through all the meeting transcripts that were done in that client's evaluation, so that we're capturing all the requirements that were discussed, the KPIs, the aspirations, so that we ensure none of that falls through the cracks as we move from client's evaluation of Acumatica Cloud ERP to the, to the implementation of it.

And then, um, using tools, um, uh, AI tools to migrate data, for example, historical data from legacy ER- ERPs into Acumatica is, is, it's been a game changer. So literally across the, the organization.

[00:38:16] Stephanie McColly: Yeah. I love it. Customer zero, that's what we call it. We, we create solutions that just solve those little pain points along the way throughout every, you know, part of the, the department.

You know, as we kind of round out the end of our session here together, we've got a couple more questions. The one I really want to hear from everybody on really quick here is, what is the one thing, an Acumatica customer is listening today, what is one thing they should be thinking about in regard to AI right now?

Andrea, let's start with you. What's one thing they should be thinking about?

[00:38:43] Andrea Hays: I would want to play the, the long game, right? So think about what your approach is to AI and what that does to your culture, your people, right? What that does to morale. So, you know, a better mindset could be how do we, how do we use AI to make our people more capable rather than are, are, am I going to be replaced, right?

And so what that means is really, really focus on the process and adoption side. I just read something that said that for every dollar you spend on development, you should spend threefold that on training, on adoption, on change management. And I think sometimes, you know, we, we can give lip service to those, but what that actually means is that, you know, the first, the person has to feel, like psychologically safe enough in their, in their employment to be able to test these new tools and adopt them, right?

So I think, I think you have to also take the, the human into account as well and their emotions, their, their willingness to try new things and, and, and trust that it's not gonna eliminate their job, right? And so I think that's kind of where it has to, has to start and end, um, in terms of a- any AI initiative.

[00:39:39] Stephanie McColly: Yeah. That's great. Armando, how about you?

[00:39:42] Armando Huerto: Yeah, I, I... It's very similar to Andrea. I think it just, it revolves around change management. Like this is happening so fast. It's overwhelming. It's overwhelming for business leaders like us, and it's equally overwhelming for our clients. And I think we all feel it, but it almost feels like the whole Keeping Up with the Joneses piece.

So if you're an Acumatica client, you know, walk, crawl, run, you know? If you just went live on Acumatica, don't expect to automate everything right away. But if you're about four years in and you have enough data and transactions in, that's a perfect opportunity to go talk to your ... And we're seeing it now where you're coming-- our, our clients are coming to us as like, "Hey, we're ready to automate this, this, and that," because they're s- they're really comfortable in the system, and it goes back to that human element.

It, it makes it less, makes it more palpable for change. So-

[00:40:33] Rob Kirkey: Mm-hmm ...

[00:40:33] Armando Huerto: you know, at the end of the day, if you have good data and ERP is running through that, I think we mentioned that before too, is that it could just be a very simple native Acumatica workflow or business event or something like that. But what I really love about AI is that it force, is forcing us to have that conversation, whereas before we didn't have that forcing function.

It was just like, "Okay, we'll get to it when we get to it." Mm-hmm. So I think that's what's really exciting.

[00:40:59] Stephanie McColly: Yeah. Agree. Agree. Rob, how about you?

[00:41:02] Rob Kirkey: I would say, you know, September is the, the month of business planning for 2027. So as clients are going through their, their SWOT analysis of, of their o- company and their opportunities, threats, uh, trends, weaknesses, I use that as the opportunity, as both Armando and Andrea have already said, to review your existing business processes, identify which ones have friction points, that have choke, choke, choke points, that have reden- redundant data entry.

And then, you know, if you haven't already started doing this, try to fix one problem as a proof of concept, uh, using AI.

[00:41:40] Stephanie McColly: Yeah. So good. So good. I, you know, when I think about it, I, I kinda think one person on your team should own the AI initiative for the organization. Obviously, it's supported by multiples, but you need one person really focused on beating the drum of consistency for it.

And if you have them go meet with their subject matter experts in each department, build a wishlist, and then start to prioritize what's most important to start looking at first, then you just create kind of a monthly check-in to say, "How's it going?" All of a sudden, you're gonna notice you start knocking out little things over time, and it makes a huge impact.

So I think all of the points are really great here today. So I wanna wrap us up here with a lightning round of just randomness. So Andrea, y- you play the French horn, which we love to talk about all the time, and you said music was kind of your first language. So what is one lesson from playing music that helps you keep your team in harmony in your firm?

[00:42:33] Armando Huerto: Oh, interesting

[00:42:34] Andrea Hays: question. Yeah. Um, I would probably say, like, the, the discipline, right? Like, you have to, if you're gonna play an instrument at a, like a high level... And I should say, like, for the record, I do not play anymore 'cause I, you know, I don't practice every day, so it's really hard to just, like, come in and, and start playing and I'll expect myself to play like I did, you know, 10 years ago or so.

But I think it's the, the discipline and the, and the continuity, right? Is that, is that you, you practice every day so that you can show up and do a really great performance, right? I think it, like the irony of the, of the AI Age is probably what, what Armando was saying earlier, right? Is that, um, what it forces us to do is, is really focus on the bread and butter of the business and make sure that we have the discipline to make sure that, you know, are we having those, you know, those pipeline discussions?

Are we having those customer check-ins, right? I think it's, I think it's getting easier for us to, to have that discipline with these tools, but also realizing how important that is as well

[00:43:28] Stephanie McColly: Great. Yeah, music lesson for all. All right. Armando, on to you. Your team said that you are a devoted Star- Star Wars fan.

Is this true? Yeah?

[00:43:38] Armando Huerto: Yes,

[00:43:39] Stephanie McColly: 100%. All right. So is, is there a lesson from this legendary, um, show that actually shows up in how you lead i-Tech support?

[00:43:49] Armando Huerto: I always tell people the li- like, in, in a joking manner, but I'm also, this is kinda how I live my life, is, "Do or do not. There is no try." It's a very famous Yoda saying, and I think I can bring that back to, to AI, is that I, I think we all feel like we, we must do or, or we're g- falling behind.

And the try thing is, is really what we're... um, I've been trying to harp to my team, is that trying means that you don't have a plan really. Doing means you have a plan. You know what you need to do, steps one, two, three. And I think if we do everything, including AI deployment and, uh, and implementation and all of that with really a good, thoughtful approach to it and, a- and doing it well, we will always get results.

So that's my favorite saying in, in Star Wars. There's many others. There's many other things, but I always, in... at work, I always, I always share that, that GIF on Teams, is, uh, "Do or do not. There is no try."

[00:44:46] Stephanie McColly: All right. Well, they, they let the cat out of the bag on that one. And speaking of that, Rob, I hear that you're called the trusted ERP Sherpa.

Is this true?

[00:44:56] Rob Kirkey: The, um, we like to look at all of our sellers as being Sherpas. We're trying to help people, you know, uh, traverse the rocky, uh, the rocky path to having successful ERP, uh, implementations, successful business mission critical application implementations, et cetera. So yes.

[00:45:15] Stephanie McColly: Yeah. So what's, um, what's one piece of advice you'd give to a customer who hasn't even started with AI at this point?

What's their first step?

[00:45:23] Rob Kirkey: I would say use all of the, the free training that's readily available out there. There's a ton of free training, uh, on the internet on how to become AI literate. Get creative. You know, I heard one of our clients said that they have a, uh, internal book of the, book club, um, where they're reviewing AI for dummies, you know?

Oh, okay. Just get in there and, um, you know, start accessing all the information that's, that's available.

[00:45:49] Stephanie McColly: Yeah. That's great advice. It sounds like a topic of a great podcast for the future too. So, um, Rob, Andrea, Armando, thank you so much for joining us and sharing practical insights our audience can use on their AI journeys.

We're all on LinkedIn. We're all active. Come connect with us. We're happy to share more along the way, and thank you, guys. Have a good one.

[00:46:06] Rob Kirkey: Thank you.

[00:46:07] Voiceover: Thanks for listening to the Acumatica ERP podcast. If you found this episode useful, please subscribe to the show on YouTube, Spotify, Apple Podcasts, or wherever you listen so you can catch every new episode.

And if you've got a minute, leave us a rating or a review. It really helps other listeners find the show. You can find every episode and more information at acumatica.com/erp-podcast. Thanks again for tuning in. We'll see you next time.