July 15, 2019
Interviewed by: David Snow
Video Clip
Login to view full video

How Automation is Transforming Business Operations

Three experts in private equity value add discuss the incredible advances offered by automation – variously identified as artificial intelligence, machine learning and robotic process automation (RPA). Private equity firms are increasingly helping their portfolio companies automate tasks that were once manual, cutting expenses and improving accuracy in the process. The experts from General Atlantic, Automation Anywhere and RSM share recent examples of automation success stories. Part 1 in the Privcap thought leadership series, “The Automation Opportunity in Private Equity.”

Three experts in private equity value add discuss the incredible advances offered by automation – variously identified as artificial intelligence, machine learning and robotic process automation (RPA). Private equity firms are increasingly helping their portfolio companies automate tasks that were once manual, cutting expenses and improving accuracy in the process. The experts from General Atlantic, Automation Anywhere and RSM share recent examples of automation success stories. Part 1 in the Privcap thought leadership series, “The Automation Opportunity in Private Equity.”

TRANSCRIPT How Automation is Transforming Business Operations Privcap: So we’re talking about a really hot topic, automation. This is an umbrella term for a lot of terms many people have heard of and might not fully understand, like Artificial Intelligence, Machine Learning, Robotic Process Automation. You’re all experts in this and are helping businesses to integrate them into their business models and business operations. And so what we’d like to learn is the opportunity for private equity firms to improve businesses using automation and automation technologies. So why don’t we start with a question for Cory from General Atlantic. How bid of a deal is this? How profound, down the road, will automation be able to change and improve businesses? Cory Eaves, General Atlantic: Well I would say, Dave, we have over a hundred companies in the portfolio at General Atlantic, and probably fair to say that all of our companies are evaluating this opportunity at one level or another. They’ve always tried to automate, but as the technology’s improved and the opportunity to automate’s improved, this has become even more important. So fair to say, definitely an important topic for virtually all of our companies. Tamas Hevizi, Automation Anywhere:  if you think about automation, we’ve been automating as a technology industry for a long time. The technology pieces are getting better and better. AI has gotten a lot better, and people are aware of it. Machine Learning, which is part of AI, has gotten better, which is the technology’s ability to figure out better strategies to hit a goal. And Robotic Process Automation is relatively new. It’s basically automating the most boring, repetitive tasks. And when these are all coming together, they can augment some of the past technologies like ERP and everything else we’ve been used to. Eaves: And I think the combination of those technologies, AI and Machine Learning and automation in general, to implement more effectively and more efficiently and drive more sustainable value in a shorter period of time. Privcap: Well, can you give me some examples of tasks at businesses that are being automated and saving the business money, and perhaps making the businesses better. Eaves: Well, I can cite one specific example, which is we had a healthcare company that was reviewing medical records, and this particular company took inbound medical records and had nurses and physicians reviewing these records. It turns out that the vast majority of those didn’t actually need to be reviewed by a human, only a small percentage needed to be. So we applied some machine learning technology to sort of filter out the low-hanging fruit in that case, and we were able to automate almost 40% of the cases that flowed through their system. That reduced labor dramatically, and it was expensive labor. It was physicians and nurses, and allowed them to focus really on the cases that they needed to focus on. And that was a great example where we bought a small company, applied this technology, and then automated a really important piece of the process. It reduced their processing cost per case by 17%, which in this business was a massive reduction in their costs. So that’s a great example I love to cite in this case. Privcap: Quick follow up question for you. I think a lot of people associate automation with reducing headcount, and that certainly happened at your portfolio company. But was there an improvement in accuracy? Eaves: There absolutely was. Having a human being review pages and pages of medical records is just prone to error, and our error rates on automation were actually lower than they were for our humans, who were doing the same task. And importantly, to your point, Dave, those humans didn’t leave the business. They actually spent more of their time focused on the cases that needed a more intense review, and the cases that flowed through the systems were actually the approvals. They were the ones that were going to get a yes anyway, and so we sort of took that step out of the system by automating that particular piece. Hevizi: So I actually start with some broad observations. I actually anticipated that a lot of the automation is going to happen to companies that are already very sophisticated in technology. I thought it’s going to be in high tech. And what we’re finding is, like Cory mentioned healthcare companies. We’re finding healthcare companies, utilities, getting into automation because they are stuck with a massive amount of manual work. So it’s a general observation, and specifically stories. We are currently working with one of the big issues in healthcare, and a lot of healthcare investors in private equity is the revenue cycle in management. Almost every portfolio company and investment thesis when it’s in healthcare is going to look at that process. And a lot of revenue cycle management optimization depends on doing something with manual paper. It’s just a lot of papers and a lot of manual steps. So we work with a lot of customers that have over 20 million claims that they have to process every year, and they have hundreds of people doing it. And there’s zero value added to that being manual versus that being automated. So that’s a good example. Eaves: Just in the accounts payable process, things that are very simple like moving an invoice through the payable process. We had one client that was processing a lot, and we did some analysis to figure out what was the cost per invoice to get out the door. At the end of the day, we were able to reduce that cost by almost $5, $4.51 to be specific, per invoice, which was hundreds of thousands of dollars worth of sustainable margin improvement on the business, which obviously equates to higher in price value, based on that. Dave Noonan, RSM: We have another healthcare company, by coincidence, that’s using RPA for exactly this, some revenue cycle management projects. In the healthcare industry, you’re dealing many times with payers, and the payers have these sort of notoriously legacy-type systems. And traditionally, the only way to interface with those systems was with a human being. You took out what you wanted to get paid for, and you had a human being type it into their website so that you actually got paid. And that’s actually a perfect example of where some of this technology can automate away the need to do that. So, great, great example. Hevizi: So I actually have another example, which I thought was good. It’s a $200 million automotive supplier that… The reason I like their story, because they weren’t set out to automate. That wasn’t what they were started to do. They were trying to figure out why their supplies weren’t shipping to them.  And it turned out they weren’t getting shipments because they couldn’t pay the invoices properly. Why? Because they couldn’t do three way matching properly. Hevizi: So they basically had a root cause analysis. This was six sigma shop, and they identified that it was basically human error. It just predictable, time and time and again, that was a judgment that was made wrong or couldn’t be made. And the vendors weren’t paid on time, therefore you couldn’t make products on time. The customer service suffered. And they ended up automating a lot of the process around it through UA match, which fixed the ultimate problem. So again, it doesn’t always start off as a cost reduction measure. It sometimes starts off as a customer service problem, that ends up being an automation project. Privcap: How fast is this technology changing and improving? I mean, what can we expect five years from now? I think a lot of people are already finding it remarkable that AI has crept into their home with things like Alexa, which I alluded to, and automatic text responses on their iPhones. As people on the frontline, how fast is this evolving? Eaves:  It’s changing day by day. I mean the healthcare example I sited at the beginning of the show, when we initially investigated and purchased that small company that provided that technology, it was crafted by 10 PhDs from Harvard, sort of hand built. Now it’s a service on Amazon, the same thing essentially that they provided on Amazon that you can buy by the minute on Amazon. And that’s in the space of three years. It went from academic research to deployment to now essentially a commodity service that you can buy online. And I think that pace just absolutely continues in the space. Hevizi: I completely agree, I think you’re going to see a lot of technology companies spring up in this space, some will be very vertical, some will be solving very nascent issues around AI. The whole race, everything is about getting better at the thinking process. So automation is useful, however there are elements of this I need to apply thinking to. Because we humans [00:35:30] think. So if I cannot do the thinking bet with a robot I still have to be around to perform those functions. So a lot of the focus and investment in this area is, how do I take that last half mile of taking some of the cognitive functions over? So if I can do 9 of 10 tasks I perform with a bot, I want to do that 10th one. And the 10th one may be some decision making. So a lot of technology’s going to evolve in that space. Privcap: Well this is a great topic, I’d love to have the three of you back to talk about it some more in the future, or at least maybe have three bots speaking on your behalf would be great. But thank you so much for being part of the Privcap Program today. 

Register now to watch this video and access all content.

It's FREE!

  • CHOOSE YOUR NEWSLETTERS:
  • I agree to the Privcap terms of use and privacy policy
  • Already a subscriber? Sign In

  • This field is for validation purposes and should be left unchanged.