AI Leadership Insights: AI for Finance and Accounting

Mike Vizard: Hey, guys. Welcome to the latest Techstrong.ai video series. I’m your host Mike Vizard. Today we’re with Alex Hagerup, who’s CEO for Vic.ai. And we’re talking about how AI will be applied to accounting and finance because a lot of those processes are fairly rote, and maybe it’s time to automate them. Alex, welcome to the show.

Alexander Hagerup: Thank you so much for having me. I’m excited to be here.

Mike Vizard: We’ve seen people talk about, say, automating a process involving invoices and trying to align those invoices with the actual payments going out the door. And a question I have, that seems like that’s the beginning of a process; how far and how deep can we go with AI these days and what should we expect tomorrow?

Alexander Hagerup: Yeah. That’s a great question. Yeah. That is the beginning of a process. It’s one of the accounting processes. The way that we’re looking at things is if you take the whole space of accounting or function of accounting and break it apart into all of the pieces that go into it, all of those pieces should happen autonomously. So that isn’t the case today, but we’re working hard to make it so.
So you mentioned invoice processing. The whole process today can run autonomously for a portion of a company’s invoice volume. So that’s about where we are today. And it’s been a long journey just to get to that point.
But I think the benefits of the development that we see in AI is that things are going to move faster and faster, even if it has taken years to get to this point. I think we’ll see a faster and faster trajectory of what can be accomplished after the initial accounting transaction processing has been performed autonomously.

Mike Vizard: We, of course, hear about generative AI every day. Is that going to be applied to accounting? And what will that look like?

Alexander Hagerup: Yeah. I think it most definitely will be applied to accounting. And there are probably too many ways for me to even know of all of them at this moment in time. But one of the obvious ones is this human interaction medium.
So there’s a lot of explanation that takes place, and gen AI can help with that. It’s good at taking instructions and understanding that and converting it to action, and then also returning answers and explaining that. And I think also, it can act very well as an assistant or a copilot for how to perform certain things.
How do you perform this accounting classification? How do you accrue this particular cost that you have? I think it can act as a copilot there as well. I think it’s going to be a central part of eventually running your entire accounting team autonomously.

Mike Vizard: And what’s been the reaction from accountants? I mean, on one side, it’s like everybody else, there’s a certain amount of… sense of fear and loathing. And on the other side of it, maybe there’s excitement because a lot of the stuff that’s being automated isn’t exactly the fun part of the job.

Alexander Hagerup: Yeah. That’s always been the case. It’s always about automating the non-fun parts of the job. Then I think one overlooked part is probably that this technology allows you to do things that you can’t do today.
So I think it’s a really good technology for enhancing the human capabilities. And that’s definitely been true earlier through other revolutions as well. And I think this is what’s happening again. The more mundane, repetitive tasks, it’s not fun and no one loves to do them. You want to be freed up to be a little more strategic. And if that’s not in your skillset, AI can actually help you become more strategic as well.

Mike Vizard: Of course, there’s a lot of regulations involving all things finance. Are the regulatory folks going to be comfortable with all this? Do they need more visibility into how all this works? What’s going to be their take, do you think?

Alexander Hagerup: That, I am not an expert in. So I’m not going to have a great answer for you there. But I think it’s clear that regulatory should have a place here, but I also think that it shouldn’t come too early so that it stops innovation.
There’s a time and a place. And at the end of the day, when you talk about accounting and auditing, the numbers have to be correct and they have to be auditable. And the numbers will always be auditable at the end of the day. You can’t audit how a human made a decision either. You can audit the result of that.
So the question is how far into that depth do you need to go in terms of auditing how an algorithm reached a conclusion versus auditing whether it was accurate or not. So I’m not an expert in the regulatory. I just think that there’s a time and a place for that, and it shouldn’t come too early as we need room to be innovative.

Mike Vizard: How hard is it to set all this up? The perception is that I need to go get a bunch of data scientists who will sit in some room somewhere and crank something out and cross your fingers and hope it works. Is this becoming more accessible? And is it just basically going to be like some sort of a SaaS application that we all access?

Alexander Hagerup: Yeah. I think it largely depends what you define as it in that question. Because it’s a very big area when you think about accounting and finance in general. But yes, to your question that it becomes more available, for sure.
You’re already there, where you have these platform models that are available basically on tap. And that has never been the case before. So it becomes more available.
And then I think you have… The reality is probably that it makes it easier to produce demos and maybe sprinkle-dust your existing SaaS application with some lightweight value. But it’s quite complex to put this into production and have LLMs run a complex business process because they are still somewhat unpredictable.
I think it’s still a very complex thing to build enterprise applications for a complex business process, even with the developments that we’ve had recently. So yeah, that’s about what I think about it right now, I think.

Mike Vizard: Every now and again, there’s a story about some sort of accounting error that led to somebody making a million-dollar mistake here or a couple of million dollars go missing over there. Will AI result in fewer of these mistakes being made?

Alexander Hagerup: Most definitely. That’s for sure. We’ve already proven that out with our technology. So we frequently hear our customers report that their error rates in the numbers drop significantly. And the AI, at least the way that we’ve designed it, knows whether it is confident about something and when it is not confident about something.
It is very, very rare that it is confident about something that is actually wrong. So you reduce your error rates quite a lot. It will likely never drop to absolute zero. There will always be some mistakes being made. But if we can make it dramatically less than the mistakes that humans are making, I think that’s a better place altogether.

Mike Vizard: Do you also think there might be less fraud? Because a lot of times, we see people inserting things into accounting processes or usually some sort of business email compromise involving an invoice or something. What is your sense of how much can we reduce fraud that, as far as I know, costs organizations billions of dollars?

Alexander Hagerup: There will most definitely be way less fraud. And AI is quite good at detecting fraud attempts versus what a human is. And in our case, when our customers are running their invoice process on Vic.ai, fraudulent invoices get detected upon ingestion because there are tiny patterns that would be different, and the AI picks up on that.
And then, of course, when you also run the classification, payment batch creation, payment execution, either executed or overseen by AI, you will have less human fraud in that process. And eventually, whether it is three years from now or five years or 10 years, whatever you believe in, when these processes run completely autonomously with very, very minimal human intervention, that fraud element should almost go away. There’s almost no reason.

Mike Vizard: How do you think the role of the accounting and finance teams is going to evolve then? Because a lot of their time is taken up doing these rote tasks. So what are they going to do with that extra time?

Alexander Hagerup: I think there are higher-level and higher-value work that still needs to be done, and maybe even more of it. I think one of the situations we see today is that there’s way too little data analysis and way too little strategy that is being performed. And I think that’s where a lot of the time should go when all of the repetitive and transaction processing work is no longer. That’s my best bet. And then you’ve got to see how it evolves.

Mike Vizard: Who’s leading the charge on putting AI into these systems? Is it the IT folks? Is it the CFO? Who’s at the forefront of this thing saying, “Yep, this is where we want to go”?

Alexander Hagerup: I see both, where some companies have really forward-leaning finance teams. It could be the CFO or a VP or a controller. And the IT team is not positive in introducing new systems into their technology stack.
And then you have the complete opposite, where the IT team is pushing for new and modernized technologies and actually helping, assisting, and advising the finance organization into making the right decisions. And obviously, that last part is the better one.
The finance team isn’t there to be software experts or IT experts. They think about the outcomes, not really what technology gets those outcomes. So I think it’s important for the IT team to be well aware of the potential technologies that are out there and how they can be securely connected to your company’s systems and be a helper to enable the finance team to accomplish more. So that’s what we like to see, of course.

Mike Vizard: So you guys have put a couple of these systems together already. What’s that one impact that people didn’t expect or that you see customers being surprised by, and how does it rock their world?

Alexander Hagerup: I think that a lot of our prospects, even though we tell and show, I think that you have to see it to believe it in many cases when things are dramatically better than what you’ve been doing for so long. You’re always skeptical about some nice pitch about something that’s going to be 10X better.
And I think that’s the thing is that we’ve been doing this for more than six years now, developing AI algorithms for accounting processes specifically, and it works. It didn’t work when we started. It didn’t work the first year, the second year. And then it started working. And then it’s been working better and better and better.
And now it operates at a really high level. And it has a significant effect on the time spent, turnaround time, fraud, like we talked about already. These areas where you just got to see it to believe it and have it in effect. There’s really a significant ROI on it, and I think that’s what surprises people the most is that it actually holds through what we’ve told them and showed them through data in the sales process.
The one thing I’ll add there, one thing I think is good is that for finance teams and maybe all other teams, but finance teams, getting started is just so important because when you’ve seen it once, you can apply that to other areas, and you can understand the impact that you can have other areas of your finance and accounting team as well.
So being able to just get started with one application, whether it’s Vic.ai for invoice processing, accounts payable process, we’ll do other accounting processes as we go along as well, or whether it’s somewhere else in your finance stack. But just getting started in tackling one process and just seeing the effects of that opens your mind to understanding where this is going three years from now.

Mike Vizard: We talk a lot about the impact that it’s going to have on jobs, and there’s a lot of “the sky is falling” conversation. But as I think it through, do you think we’ll soon get to a point where if I’m a finance or accounting professional, do I really want to work for an organization that isn’t using AI to do the job? Because otherwise, the job is just going to be kind of… I don’t want to say miserable, but it’s certainly not as much fun as it could be.

Alexander Hagerup: Yeah. I think we’re headed there. So I guess you have the two camps. You always have the people that are holding on to the old processes, and they feel safe and secure. And then you have the newer generation and the forward-thinking people that understand that technology is an enabler.
I always think the fear of jobs is always over-exaggerated, and it’s been the same way through history, where there’s always some new technology coming along. And you think all of the jobs are disappearing. And then some of them do disappear. People get elevated or they do other things. There are new jobs created.
So I think that’s the case as well with AI and what’s going to happen over the next three to 10 years. There’s going to be jobs we can’t even think about right now that are going to be created. And they’re going to be maybe much more fun than what it is to enter transactions into a computer all day. I’m obviously an optimist when it comes to all of this, but I think history has proved that out time and time again.

Mike Vizard: All right, folks. Well, you heard it here. Look, the AI genie is out of the bottle and it’s not going back in, so you might as well enjoy the ride and see where it goes.

Alexander Hagerup: Exactly.

Mike Vizard: Alex, thanks for being on the show.

Alexander Hagerup: Appreciate it. Thank you so much, Michael.

Mike Vizard: All right. Thank you all for watching this latest episode of the Techstrong.ai video series. You can find this one and others on the Techstrong.ai site. And once again, thank you for spending time with us. Until we see you again next time.