Future-Ready F&A: Hyperbots Invoice Processing Co-Pilot

Brian Kalish:
Good morning. Good afternoon. Good evening. Wherever you may be in the world. It's great. We live in a world where we've got tremendous technology that we can reach the globe all at once. I'm Brian Kalish. I will be your interviewer and moderator for today. The topic that we have is “Hyperbots Invoice Processing Co-Pilot”. We're very fortunate today to have Dave Sackett joining us.

Again, Dave, thank you so much for joining us. Just a bit of background – I'm Brian Kalish. I've been involved in FP & A and in Finance and Investor Relations & Treasury, for the past 30 years. I've been working with HyperBots for the last 2 years. 

Dave, if I may do an introduction of yourself – Dave is a finance leader with expertise in AI ERP systems and Operational Finance. He currently serves as the Vice President of Finance at Persimmon Technology, and has held a number of senior roles at companies such as ULVAC Technologies, Visibility Corporation, and Nova Biotechnical. In addition to his corporate roles, he has also co-founded AI One and advises multiple AI focused organizations. So Dave, thanks for joining us today.

Dave Sackett:
Yeah, thanks, Brian. Thanks for the introduction.

Brian Kalish:
Absolutely well, really. So today, our audience really wants to understand how Hyperbots Invoice Processing Co-Pilot is redefining the month end processes for financial teams. 

So with that, let's really kind of jump into it. The 1st question that I was thinking about Dave, is what challenges still haunt Invoice processing. 

Just to share a little bit of what I'm hearing, challenges in invoice processing include inaccurate invoices laid, delayed payments, time consuming manual processes, scalability issues and limitations with back office or accounting systems integration. What's really important is, what are you seeing?

Dave Sackett:
Yeah, so those are pretty common. We also have price differences on Pos. We have 3-way matching where something may not be received or in the incorrect line. Like issues with, not including things over billing, under billing; so many things that you would have to have someone really focused on AP, to make sure we didn't post the wrong information or paid someone who we shouldn't have paid.

Brian Kalish:
So kind of right to the gist of it. How does Hyperbots Invoice Processing Co-Pilot really work, from an end to end perspective?

Dave Sackett:
Yep. So what we're doing is  robotic process automation with AI supporting it, meaning we'll have invoices sent from an email address to Hyperbots that AI technology is going to read. Then it looks at our system, and matches it up, making sure that everything is accounted for, and that we're taking discounts if we have to for doing all that. It checks like a bot would, to make sure it follows our policy and it looks for that receipt. It's looking for the matching, ensuring the price and the Po is correct.

Then it puts it into HyperBots as if it was a person typing in an invoice. And then it's up to us now to go in and see what HyperBots did and ask – Did it do it correctly? What do we have to change? Is there an exception? 

Going through this process, we’re teaching HyperBots, our agent, about what our invoices look like, how we'll pull data over time. As we use this tool, it'll be more efficient and we’ll be able to streamline it so that we don't have to double check. It just happens and invoices just get pushed right into the system.

Brian Kalish:
Great. I can see so many benefits from this. What hard benefits have you been able to measure?

Dave Sackett:
So when I look at the ROI on what we're doing, it’s essentially the overall replacement of a person. We used to have two people on accounts payable. Once one person got terminated, I had the option to either hire a person or use technology. 

So by choosing technology to do all that data entry that the person was doing, I can do that much cheaper with Hyperbots. Paying as you go versus paying someone's salary to do the same exact thing – the beauty of it is that now, when we scale and grow, when I have more and more accounts payable to process, it will be technology taking up that slack, not the need for hiring more headcount or bodies in my department.

Brian Kalish:
Oh, we really see that across the board. And again, curious about the current situation : Is the person happier? Because I would assume a lot of what I would consider low IQ activities are off their plate so that they can actually spend their time more efficiently and effectively.

Dave Sackett:
Yes. So yeah, our Accounts Payable Supervisor has a very busy role generally, and to have her time keying in manual invoices doesn't make sense. So she's excited that now a bot will be doing that work, and she'll be managing the bots output. And it is solving problems and also being the communication on the human side of AP. So, she's excited for it.

Brian Kalish:
Excellent, the other comment that I get all the time is the fact that – and you commented on it – what I call an electronic employee, right. When you think about it –  I don't know if you've seen this, or you've dealt with the issue. But from a security perspective, people are like – do you have a human doing it today?

Dave Sackett:
Right. It's absolutely on you. You have to onboard employees. You have to make sure that they follow rules, and they don't go where they're not supposed to go. Same thing with Hyperbots, in terms of an agent that we're using to do this same data entry function that happens all electronically. So it's very much the same thing. You need to make sure that when you're onboarding Hyperbots, you've got an eye on cybersecurity and company policy, because it is really like having a digital employee now contributing to the finance department.

Brian Kalish:
Yeah, and if I may – without having HR issues.

Dave Sackett:
Yeah, they can work 24/7, and don't have sick days. So yeah, no fights with other departments, and a lot of headaches go away.

Brian Kalish:
Absolutely, absolutely. I don't know if this is a fair question, but what single capability has unlocked the most value from your perspective?

Dave Sackett:
Just from a pure data entry perspective – pushing through the automated invoice system, being able to take any invoice that gets invoiced to a specific email address and then pushing that from the bot right into our ERP system saves a ton of time and as well as accuracy. Like there's no missed keys, and there's no missing freight on an invoice. The bot can be very thorough and categorize that invoice and take advantage of even 2% net, 10 or any other kind of payment terms that other people—it's just not on their radar. But you program the bot to have all the rules, so it's constantly looking for being compliant with whatever rules you assign it.

Brian Kalish:
Yeah, by definition, it is compliant, right? Because it doesn't—

Dave Sackett:
Yeah, it doesn't choose not to work.

Brian Kalish:
It doesn’t have a bad day and miss things, right? Again, it's that whole idea. Some people would say it's additive. I think it's multiplicative by adding agents because, again, this work gets done and the rework—going back and finding problems that humans did—is now eliminated.

One of the things that I've seen with Hyperbots is this dynamic GL coding. I don't know if you've been involved with it too much, but to the extent you have, how reliable have you found that to be?

Dave Sackett:
Very reliable. Yeah, it's just data. They're assigned to picking it up off the PO—our purchase orders. So we're very much purchase order-driven, as a manufacturer. So from that account number structure, it's very seamless in terms of going into the right GL account.

Brian Kalish:
Great. If I may touch on another function, it's the autonomous exception, which is interesting, right? Because we're moving from automated to autonomous in finance. Gartner should give me a commission for that statement. But how does autonomous exception handling actually reduce the workload?

Dave Sackett:
Yeah, the bot takes it on. It's got goals. It wants to process everything it gets. It can take action. It can push into the ERP system if everything's fine. If something's not fine, it can put it into a “waiting to be reviewed” spot in Hyperbots. So there's so much more kind of intelligence built into processing AP that didn't exist before—or at least not in my solutions—that we're taking advantage of.

Brian Kalish:
We made reference to people being concerned – new technology. Again, I have my opinions—feel free to push back. In finance, we're small-C conservatives, right? So technology is a little scary for us. You and I, I think, if I may be so bold, are finance people that know a little bit about technology. But we realize that's not our strength. So again, I think there's always just going to be a concern, as we're the fiscal officers of the organization. What have you seen? What safeguards exist for errors and outages?

Dave Sackett:
Yep. So a safeguard that we have in place is human review. We are ultimately—we have to be the certifiers of the truth. So we have to review everything Hyperbots does, at least at this point, to make sure things are going where they're supposed to. Everything is accounted for. Right now, we're still in the implementation phase. I'm going to expect over time to give the bot way more freedom to push it in, just like a normal employee would.

That’s the goal for what I have in terms of automation. Remember, my very first job out of college was a cash accountant, and they handed me a ledger book. “Here you go, dude. Here's your pen. Keep track of daily cash with this book.” And I'm like, I just graduated college. We used lotus 1, 2, 3. I could make this a spreadsheet. And we're golden. So like my whole career has been trying to take something that was done manually and automating it. So automating it, using excel, using Hyperbots - my whole idea is to try to streamline and do things more, faster, efficiently, and a better way leveraging technology.

So I've done it kind of my whole career and had success with technology. And that's kind of where I am today. I'm a blend between maybe a technology person and a finance person.

Brian Kalish:
Yeah. Well, I think we have to be. That's what's changed. Right? You really can't do the job, although I have to admit. This may have been even before you. My 1st version of Lotus 1, 2, 3 - we double loaded with dual drive, with 5 and a quarters, and you do your work, and then you'd pull it off. 

So yeah, we looked like geniuses to your point. The first job I had, was very much similar, right? And I do see that progression right? And then in where we are now, when we start talking about Hyperbots AI, there was a point where you went – “Okay, how many times am I doing this?” These repetitive strokes that you realised you needed to build the macro right? So there's an investment.

But with a much greater ROI on the backside. Then we got into RPA. Now, I think we're really at that point where I'm a huge believer. I'm way out the curve. I think we're living in the Industrial Revolution in real time like this 30 year battle that we've had, where our folks spend 80% of their time on data, acquisition, verification, reconciliation; and 20% of the time on the analysis, I think we can turn that paradigm on its head.

Dave Sackett:
I agree. No, and I'm a hundred percent in agreement with that, especially with Agentic AI. Where now you've got agents that can do so much more than simple machine learning.

Brian Kalish:
Yeah, it's really incredible. And again, that’s why I think in all honesty, having sessions like this are important, because the way I put it in the office of finance. We already work 26 HA a day, 8 days a week, 60 weeks a year.

And we have to know technology now, because AI is not going to come to us with these solutions because they don't know the problems we're trying to solve for so it really, we need to be the champions. If we want to make our world better, we have to, spend the capital and make it better.

Dave sackett:
Agreed. And especially people in finance — we have certain pain points. And me – it's hammering in accounts. Payable manually doesn't have to happen and doesn't have to be so. I knew there was another alternative out there. So I was motivated in finance. We have to make this upgrade. It's probably not anyone else's radar, because it's not affecting them. But it is affecting my department.

Brian Kalish:
Efficiently enjoy even in enjoyment. Right? I know I would be very keen, to be coming into an organization that's leveraging an agentic AI to solve these problems that you go over rather than like you said, you're going to be – checking excel spreadsheets.

From a drawing talent perspective, I know you mentioned. You're in the integration. How has that gone so far? So the way I kind of think I had originally thought, like how quick was the ERPintegration. I don't know if you're through that point yet or not, or just how the process has been so far.

Dave Sackett:
No, the process has been good. Everything's kind of methodical, going to plan. Everything's mapped out –  People involved to make sure we could do the Apis to connect Hyperbots. And we have an on premise ERP. So they're a cloud system, and going from cloud to Prem ERP is squared away. They've got demo accounts, or our test database, and on both sides. Hyperbots and the Persimmon side have been putting time and effort into it to make this project go forward. And yeah, I'm thrilled with the progress that we've made and how it's working so far.

Brian Kalish:
That's great. So you went through, again making assumptions – So obviously, there was a process that you went through as far as evaluating potential AP tools. At the end of the day, if you want Hyperbots, the big question, I think, for the audience is, why?

Dave Sackett:
Yeah, a couple of major points in terms of why I picked hyperbots. One, I have to like the people I do business with, so it doesn't matter if you're some big company that has a great solution, but I'm just a number to you, I'm not interested. People that can show me, “Hey? We're going to be your partner. We're not just a supplier. We're going to actually work with you and help you make your business better.” So it's that personal relationship with Hyperbots and the people. 

Second, it's the cost from my finance hat. Now like, is this a crazy dollar situation? No, it's not. It's very affordable, and it pays as you go. So if I only have the bot working a little bit, I'm only paying transactionally a little bit.

 I've been involved in other RPA implementations where they wanted a huge upfront cost. And then not only that upfront cost, but now you've got maintenance on top of that. So from an ROI point of view, it's much longer and more cash outlay in kind of that traditional model that I'm used to, and that's what I find attractive about Hyperbots. It's that if your bot's only doing a little bit of work, then you're only paying a little bit. If you now have to grow 3 times, and now you're swamping. The bot would work well, now you're paying more, but that's still way less than having to hire all new people to do it. So from a scaling point of view, I know my company is growing, and we're going to have more and more transactions.

 So it's only natural now to get in at this stage, introduce Hyperbots that can now grow with us, so we'll take on more functionality. We'll have the bot doing more and more transactions, and really leveraging the technology to gain productivity and efficiency.

Brian Kalish:
That's outstanding. Yeah, the number I kind of hear is about, about 80%, right? That's kind of what it's almost, it's the old Pareto rule. Right? The answer to every question is always 80%. But it seems like, but I think that's important for people to understand is that you're not going to just have this. Again, if we keep in the context of an employee, you have an employee that knows everything right. But if they know 80% of it, then, like you were saying the structure you have, then that 20% that it can't handle straight through gets kicked up. And just from an efficiency and effective perspective, I've got to think that's tremendous.

Dave Sackett:
It is because, yeah, your regular people will make mistakes. No problem. A bot will, too. And the more that the bot sees the invoice processes, we train it exactly – “Oh, you missed this. This is how we do this one. This is how we do that one.” Any of those exceptions, we can now train the bot to not have them next time. So they're continuously learning, just like an employee would learn to do their job better. Same with AI, same with Hyperbots. So it's that going through and really working with the tool to try to maximize what a bot can do versus something that gets put into exceptions.

Brian Kalish:
Yeah, as we get ready to close here. The only thing I would add to that is remembering when we teach a human. Then the human knows when we teach AI. Then all the bots know. Right? That's pretty powerful.

Dave Sackett:
It is.

Brian Kalish:
So, I didn't know if you had a thought on that.

Dave Sackett:
Yeah, as we roll out other bots. We've already got that linkage now between Hyperbots and our ERP system, Epicor. So there's access to the data. So now, if I need to branch out, other people in my company need to use AI with their data that it's in Epicor. That link is there. That solution is there for the taking. And so now, rolling that out will be easy, because I already have a vendor. I trust I've got this solution already proven in accounts payable that it works. And so now to introduce it to other departments, now everyone's leveraging this solution. That's efficient and affordable.

Brian Kalish:
That's really incredible. Well, your experience really shows why the market, as we were talking, is really shifting from what we would call in a pre-pandemic world. That's where I kind of kind of call what time is for me. When we were talking so much about automation to an environment that's truly leveraging. Agentic AI. So really, Dave, thank you so much for sharing your insights on Hyperbots Invoice Processing Co-Pilot again. Thank you to the audience, and wishing everyone a good rest of your day.

Dave Sackett:
Great, thanks, Brian.

Brian Kalish:
Pleasure.