Matt Hardwick: All right, let’s go ahead and get started. Thank you for joining, everyone, today. Today is part two of our webinar series around invoice processing and how TurnPoint Services has modernized invoice processing. I’m here today with Paul Barth and Jason Pierce from TurnPoint Services, and we’ll just get started. Paul, Jason, if you don’t mind giving yourselves an introduction and a little bit about your role at TurnPoint Services.
Paul Barth: Jason, go ahead, you’re up first.
Jason Pierce: Okay, perfect, yeah. So, I am the VP of Finance and Integrations and Systems at TurnPoint. I’ve been at TurnPoint for five years now, so essentially anything that connects to TurnPoint’s finance tech stack, which we use Intact, basically goes through me. And any acquisition that we have, I go on-site and I kind of onboard the accounting teams and all of their financials into our software systems.
Paul Barth: Awesome. Thank you, Jason. This is Paul Barth. I lead our Digital Acceleration Initiative at TurnPoint Services, which is where we use emerging technology to generate leverage for our employees to improve the customer experience or drive back-office productivity. I’ve been with TurnPoint for about a year and a half, and while I live in Madison, Wisconsin, I’m on a business trip right now, coming to you from sunny Palm Beach, Florida.
Matt Hardwick: Oh, nice, very nice. Getting some sun for the first time in a while. So I guess we’ll start in just… how does invoice processing fit into your team’s day-to-day responsibilities? I know, Jason, it probably affects you and your role a little bit more, but could you just give us some insight into how that process in particular affects your teams?
Jason Pierce: Yeah, so we’re comprised of about 60 companies. We process anywhere from 55,000 to 60,000 bills a month. We have POs for about 60% of those bills. We do a lot of job costing. And traditionally, it was your traditional AP workflow, essentially where we have a large team of AP folks that sit in inboxes for those specific brands. Bills land in those inboxes. They have to open every single one of those emails, review the attachments, enter the bills, or manually match those bills to POs in Intact. Completely manual process. Intact did offer some automated solutions, but nothing that would allow us to not require every single email to be touched, every single bill that we receive to have to be at least looked at by someone. So pretty—if you can imagine—55,000 bills, that’s a lot of time and people that we have to put towards processing AP.
Matt Hardwick: And was this something that there were folks at each brand doing? Was it a shared services model, a combination? How was that being handled manually?
Jason Pierce: Yeah, it’s a combination of the two. We do have a shared services model for brands that maybe aren’t big enough to support having a dedicated AP person, but for over half of our brands, we have basically on-site AP people that are processing bills or matching POs.
Matt Hardwick: Right, so this is a massive time spent for folks. Was the time the main challenge here, or were there other challenges that you were seeing with the process that led you down the path to automation?
Jason Pierce: Yeah, so Intact is limited as far as speed. A lot of times our AP folks were watching a screen spin as they were processing bills, so a lot of wasted time there. And so, if you think about it from a transactional perspective, having to go in and attach things and save things, even if they’re efficient at entering bills, it still takes time to process each invoice.
Matt Hardwick: Right, right. Was there a component of error mitigation as well when you were looking at automation, or was it really just that time spent?
Jason Pierce: Yes, as far as human nature is concerned, people will miss key things. They will enter the wrong vendor name, receive the wrong item on POs, so there is a human element of errors that you would expect to see with 55,000 bills being processed in a month.
Matt Hardwick: Yeah, yeah, exactly. So you’re doing this process manually—a lot of folks spending a lot of time doing this—and still seeing some errors, but really it’s a lot of that time spent, right? So was there a specific moment or sort of a realization where we decided that, hey, automation is going to be a better fit for this process?
Jason Pierce: Yeah, I think as we looked at our brand ecosystem, they’re all leveraging the same vendors for the most part. We’re buying the same things—we’re just doing it across 36 different states. And so, as we grow, because we do buy businesses, we were looking to have a way to not only grow and scale, but try to scale without adding people.
And so, in doing this, AP obviously being a significant portion of our finance team, we were looking for ways to automate everything that they do, as well as add accuracy and speed up exception handling as well. That’s really the motivation for us automating our AP process.
Matt Hardwick: And then I know there was some collaboration with leadership around how we wanted to go about doing this. Paul, maybe this is a question for you—could you speak a little bit about how the leadership team played a part in evaluating this process for automation and getting us to the point of implementation?
Paul Barth: So, if we take a step back and acknowledge that we inserted this project, this opportunity, this business problem, into a pipeline of ideas—and I don’t know if we’re at a place that we want to share any screens or if Bella’s ready to do that—but I put a couple things together to kind of set the stage of how we got to where we are.
And what it required was really an investment by the leadership team in a program—a use case in private equity speak—where we placed some small bets early to learn where we could automate. And again, thinking about customer experience, thinking about revenue, thinking about growth, but also thinking about back office.
And we’d actually started over a year ago on some smaller opportunities so that we could learn to prove that automation was an important component for the initiative that I lead, and part of really—think of it less as starting a project and more building a movement.
So how do we get the executive team, the sponsorship of the work that we need to do, excited in seeing automation as a capability that we want to grow in the organization—and that is beneficial to them over the long term?
So we built a pipeline of ideas, we’ve structured a program around it. Maybe, Bella, if you want to go to the next slide—I’m just kind of speaking in generalities around sort of what we’re looking at here—but it is really structuring this so that there’s a series of investments that are going on, a number of activities that we have going on, some large, some small. I think there’s a third screen, too.
I don’t need to speak to specifically what’s going on here, but it’s really to reflect that the invoice processing—the stage was set for invoice processing to get going quickly by the year of work that preceded it to get the executive team aligned and bought in to what is automation as a capability, what is the investment, and who is the right partner to do it with.
So by the time we got to the automation piece—and, alright, thank you, Bella, I just wanted to maybe pull those up for a second—that was the foundation upon which we got to here.
So then now, in cost-benefit analysis—or in private equity parlance, EBITDA—it’s all about EBITDA: earnings before interest, taxes, depreciation, and amortization. Who doesn’t love it?
We had a framework to have that conversation. And when we looked at what the level of effort was to automate this work, the cost of rework, and then together with Lydonia, doing the technology exploration to identify what is the best way to do it—now we’ve gotten to a cost.
We have a novel set of capabilities—the agentic, what is it, the generative document intelligence that we’re using was pretty novel for us—and it married up with the RPA work that we’d already done.
So that we had a pretty high confidence of what this was, what it was going to cost, and on the other side, we were able to articulate discreetly where we thought the savings were going to come and how we were going to achieve them.
So when you can mitigate the technology risk, you have a high level of confidence in what you’re going to do, how you’re going to do it, who you’re going to do it with, and you have a set of KPIs that you’re going to measure on where the benefits are going to come from, everybody’s got a high sense of trust, they have an understanding of how you’re evaluating this investment against the pipeline, and it made it very easy.
This boiled right up to the top, where it was like, we’ve got to go. So that’s probably a longer answer than what you were looking for, but that’s how we got here.
Matt Hardwick: No, that’s great, and I think it’s important to understand that background of the work being done to get to that point, where a lot of the evangelizing of it is, you’ve got to see it to believe it for some people, right? So it’s a combination of that and having the conversations in the background and building out that pipeline. I know we did a lot of work with different business units on just evaluating ideas, building out that pipeline to be able to have something like this rise to the top.
Paul Barth: Yeah, and the thing I would add to that is it’s about alignment, and it’s about, you know, in my role, I’m a bit of a lobbyist in the organization, right? I don’t own finance, I don’t lead IT, but I need the CIO and the CFO to be very excited about this. So ultimately, that needs the CEO to be bought in, and that’s really where the sponsorship has to start.
And that manifests by automation being one of the top two priorities of our initiative. So that gives it priority for the organization, so it isn’t just another project in the funnel—it’s one of the key projects in the funnel.
And the good news and bad news about that is you get the resources you need when you need them, but then you need to deliver. But we are very much on track to achieve that here, so it’s a great story.
Matt Hardwick: Yeah, it really is. And Jason, just to go a little bit more tactically on this, could you walk us through a little bit of how Lydonia and TurnPoint work together to support this implementation for invoice processing specifically?
Jason Pierce: Yeah, so essentially, Lydonia did not assume anything. And so we basically said, let’s look at what automating AP looks like. Let’s look at what automating the conversion of purchase orders looked like. And so we did an extensive walkthrough of every click, every document, where those documents come in, who they’re coming from, and then what key stakeholders, if there are exceptions, need to be involved in the resolution of that.
Lydonia provided us with a lot of guidance on best practices from an RPA perspective, since it’s new to us. And so it made the process very, very easy. And so as we were deploying our first bots, we weren’t just starting at kindergarten, they were actually working.
When we were, you know, we would do what we would call smoke testing, where we would target specific bills, see how they go into the system from the RPA perspective, and then review those results until we had confidence that the bot was processing things in the manner that we wanted it to. Any tweaks or changes, the Lydonia team were able to action those very quickly.
Matt Hardwick: Excellent, yeah, that’s great to hear. And then as far as that generative document intelligence tool that Paul mentioned before—had your team ever, or I guess I should ask, in your past lives, Jason, had you ever operated with something like that technology? And could you give us a little insight from your end, from the business side, how you thought that product would work, and kind of how it worked once you saw it in action?
Jason Pierce: Yeah, so not really. OCR is really the only imaging document extraction tool that I’ve ever used. With the document extraction tool that Lydonia uses, my initial thought would be it was going to be very rigid—that if you have 1,000 different invoice formats, the extraction tool is not going to be able to extract the data that we need to process.
But that was not the case. We did some training with some of our larger vendors, but for the most part, all the data that we wished to extract from the bill, we were able to extract. We’re able to put that data into our accounting system.
And kind of as a secondary result of reviewing what that data looks like, we also decided to create kind of a purchasing database, so we know exactly what we’re purchasing, how much we’re purchasing, the price that we’re paying, and so we can hold our vendors accountable for what we’re buying. And we can look at exactly what we’re buying across our entire platform.
Vendors may be charging us one price for one thing in one place and a dramatically different price for the same thing at another place. And so this gives us that data to analyze our spend, and it’s all in one place. And that was kind of something we did not realize would be useful for us, or would be possible, until after we kind of got into the project.
Matt Hardwick: Yeah, I think that’s important to call out as sort of an unintended benefit of automation here—allowing you to have just that larger data set across all of your 60 brands, to be able to say, hey, we’re being maybe charged differently here in one brand than another, and so on.
So I think seeing some of those downstream, more unintended benefits that can pop up is great to hear that from you folks. But how did some of the changes that we did expect—how did those take place and take hold within TurnPoint once this process was automated?
Jason Pierce: Yeah, essentially, we’re exception handling now. We’re not having to touch every single email that comes in that contains a vendor bill. If the bot has low confidence or a question in something specifically, it escalates that.
But as of now, once we’ve trained the bot and have resolved data issues that are on our side from a vendor naming perspective, the bot is pretty much handling those bills now with about an 85% to 90% success rate.
So if you look at 55,000 bills, if we’re only having to touch 5,000 of those now, and that’s more so answering questions—the bot may have low confidence in something, or it may be a difficult invoice to read for a new vendor—not having to touch those 50,000 bills now allows our team to look at other things, more valuable things, improve our job costing, instead of just data entry and bill entry is concerned.
Matt Hardwick: Yeah, that’s excellent to hear. And just for context for the folks on the call, that 55,000 number—that’s a monthly number, correct?
Jason Pierce: Correct, yes. Yep.
Matt Hardwick: Yep, so just to give an understanding of the scale of this process and how this process can scale up and not lose any accuracy or straight-through processing.
And then as far as some of the things that we’re doing around the invoice processing, I know we’re taking some steps to scale the process out and add increased capabilities. Do you mind speaking on, Jason, some of the maybe Day 2 enhancements, or some of the things we’ve done since that initial deployment, and how those have impacted the process as well?
Jason Pierce: yeah, absolutely. So we now have a bot that will—We have certain brands that have their purchasing process is backwards, where the bill essentially triggers the creation of the PO.
And so basically, we have a bot that’s already extracting all the data that we need from that bill, and so now we are actually taking that bill when it lands on the AP inbox and creating the purchase order for that using the bill.
And so the bot’s doing that today for the brands that fall into that purchasing workflow. What that allows us to do is we get on-time job costing. At month end, we don’t have what I call a 10-car pileup, where we have 50 vendor bills from a vendor but no purchase order because the purchasing team never created that PO, which then delays our month-end close process.
And so we’re able to now create those POs, create them on the job, we get accurate job costing. It allows us to automate our entire platform of companies. We get improved job costing, we get exception handling earlier in the month as opposed to at month end.
And we can actually reduce our close cycle because we’re no longer waiting on AP to get all the bills entered at month end. So we know where we are from a financial perspective at the brand level a lot sooner today than we did prior to this AP automation project.
Matt Hardwick: Yeah, that’s great, and I know we’re still in progress with some of that work to be able to scale it out to all 60 brands, given the different way they operate in the process, but that generative intelligence, generative document intelligence tool, and the ability to extract all the key information up front, that spans across all 60 brands. It’s just some of the picking and choosing of some of the things upstream or downstream from that process.
Jason Pierce: Yeah, I think it illustrates the flexibility that the automation tool has, where it can accommodate multiple AP workflows. So, if you are an aggregated business with different purchasing workflows, you have the ability to align each segment of your business with defined workflows that meet their needs. And that’s one of the things that, when we kicked this project off, I only thought 20 of our brands would be eligible for automation. But as we explored the capabilities of the extraction tool, we realized that we could build workflows across our entire platform.
Matt Hardwick: Excellent. And then as far as continuing to scale that out, and this’ll lead into some more kind of high-level TurnPoint priorities, but what other—are there any other places where we haven’t captured automation quite yet that we’re thinking about getting into as far as invoice processing is concerned, or are we at the point where we’ve covered most or all of the spectrum of the ways that folks are invoicing today?
Jason Pierce: Yeah, as of right now, we have pretty much checked the box on invoice processing for all of our brands. I think where we could go in the future is creating POs based on a price book, so extracting the items from a bill and actually creating price books for our brands, and identifying and actually building a PO for the exact items that we’re buying that’s on that bill. And so that’s something that we’re going to be exploring. And again, it goes down to accurate job costing for us. We do over a million jobs a year, and so that’s important for us to manage our businesses. And so we have PO creation today. But going down the line, we would have PO creation creating POs with the specific items from our price books that we’re buying. So that’s something we’ve discussed and is on the roadmap.
Matt Hardwick: Excellent. Yeah, we’re looking forward to seeing that solution come together. And then looking forward, and more of a wider-ranging TurnPoint, at a higher level, Paul, could you speak to some of TurnPoint’s priorities with automation for 2026?
Paul Barth: Sure. To this point, I think a lot of our, I’ll call it the first wave of investment, was really looking at things like the back office work that we just talked about, where we can automate, we can eliminate some of the manual work that is just hard to do, prone to errors, and not really a value add, to now we’ve made it an accelerant to our business, to where we’re an acquisitive company, so as we purchase companies, this becomes part of the sell, so that’s important.
But now we’re thinking about where do we use automation to grow our business. So, how do we get closer to our customer? How do we handle our leads faster? How do we do things using automation that are, again, growing our business organically? So there’s a lot of investment in different ways that we can either get leads or manage our leads and convert those leads using automation. And I see that as probably the big new thing in ’26. But I do believe there’s still quite a backlog of ideas that we have that are also continuing what we just talked about in the invoice space, where it is more back-office automation productivity as well. So I do see us kind of concurrently going after the work that we see in front of us, but then expanding it into the growth side.
Matt Hardwick: Absolutely, and in your opinion, how do you feel Lydonia fits into that kind of spectrum for you?
Paul Barth: It’s a combination of a delivery partner, so that when we—it really—they help—you guys helped us in thinking about how to build a funnel, how to build a pipeline, how to identify the opportunities.
And then it has also been bringing solutions to the table for understanding the art of the possible. I go back to GDI, like, I don’t think that was something that was on our radar. I hadn’t used that before, Jason mentioned he had not. And that’s where you guys brought something that allowed us to see something that would be very different than what was even on our radar prior to the conversation. So it’s that combination of strategic prioritization, execution and delivery, roadmap development, and then, again, that last piece is the art of the possible.
Matt Hardwick: Absolutely, yeah, I know we’ve built a great partnership over the last, I guess, 18, 24 months, and we look forward to keeping that going. Any other final comments? I know we have a couple questions that came in in the chat, but any other final comments on whether it be the invoice processing process itself, or Turnpoint’s journey with automation and AI? Anything from either of you?
Paul Barth: The thing I’ll start with is some of our most promising work that we’re doing right now started in more of a discovery mode. So we didn’t know, on some of this lead stuff, where it was going to go, but we thought there was something there, and we started small, and we stuck with it, and it turned out to be just a tremendous bet that paid off.
I wish my 401K paid off like that bet. And my counsel is, you want to balance putting some experiments, like having a little—it takes some executive support, it takes some courage, and there’s some risk.
But without having an ability to do some experimentation, we probably wouldn’t be as far as we were. And if we waited until some of these things are quote-unquote fully baked, you’re probably late to the party, and your competition is probably ahead of you at that point. So there’s a—we’re not a build-it-and-they-will-come company, but at the same time, you’ve got to play some strategic bets. That would be the one piece that I would provide guidance to others on the macro side. Jason, I don’t know if there’s anything from your side.
Jason Pierce: Yeah, so really from a—as we standardize processes, we really discovered we had broken processes upstream. And so when you define what your workflows are, and you do that defining with the automation tools that you build, you then discover that people are not doing the processes upstream properly.
So it’s really given us an opportunity to see what those upstream breaks in processes are, and then standardize what we want our vision to be, and what we call the TurnPoint way, upstream to fix our purchasing and to fix our purchase-to-pay processes.
And so as we defined and automated through RPA, it allowed us then to identify upstream issues that we’ve been able to fix, and really ultimately have a standardized business practice across our portfolio.
Matt Hardwick: Yeah, I think that’s a great point, Jason, and I think it’s something that we’ve seen at multiple different places in TurnPoint, and we see with other clients as well, is automation and AI become a great tool for process optimization. It really shines a light on some of those cracks in the foundation, so to speak, and allows you to fill those gaps, whether it be with automation or process re-engineering, that allows you to do automation more seamlessly and integrated throughout your process. So I think that’s a really great callout there.
So Paul, Jason, thank you so much for hopping on with me and breaking down really your program at a high level, and also the invoice processing process in general, answering all the questions. You guys have been great partners to Lydonia, and we look forward to doing more and more work with you all in the future.
Jason Pierce: Thank you, Matt.
Paul Barth: Yeah, thank you, Matt. Appreciate it.
Matt Hardwick: Awesome. And thank you to everyone who joined. Bye-bye.
