RSM: Reaping the benefits of RPA

Machine learning solutions can play a key role in any firm’s value creation arsenal, with a minimum investment of time and resources, says RSM’s Dave Noonan

This article is sponsored by RSM

A great deal of the discussion surrounding artificial intelligence in private equity concerns its revolutionary potential, and without question that can be exciting and worth examining. But that focus can serve as a distraction from what machine learning solutions can do right now for GPs and their portfolio investments.

Those solutions might not be smart enough to replace the deal team just yet, but their ability to automate and continuously improve processes and procedures can have a real impact on the bottom line at portfolio companies.

We sat down with the head of RSM’s private equity consulting practice, Dave Noonan, to discuss what Robotic Process Automation solutions can provide the industry today.

How should GPs be thinking about RPA solutions?

Dave Noonan

RPA solutions are another lever of value creation GPs can pull, and one they use at various times during the investment cycle. Whether they’re doing a carve-out or an add-on acquisition, or merely looking for synergies at a given target or platform, RPA solutions can drive additional cost out of the business.

In my experience, private equity firms have done a stellar job at driving operational efficiencies through various process improvements, but RPA solutions and related technologies can create even more efficiencies, and with an AI component, these technologies continue to improve all by themselves.

What kind of processes can these RPA technologies best automate and improve?

Any processes that are largely manually driven, that are frequently executed and require access to more than one system to complete. Any process where there are multiple iterations, such as new employee onboarding, where a new worker has to set up their ID and desktop interface so they can access all the relevant systems.

Things like accounts payable and accounts receivable, where we see a ton of activity because it’s usually manually driven: an invoice hits someone’s desk and has to be recorded in one system and approved through another. This process is time-consuming, repetitive and easy to automate.

So RPA solutions automate these functions, but beyond that, how do they improve these processes?

The digital bots doing the work now aren’t like their human predecessors; they don’t take coffee breaks or leave at six to go home. They work 24/7, 365 days a year. And the AI component learns how to execute this process faster and more efficiently as it goes along. But there are limits here. If the company has a bad process at the beginning, AI won’t necessarily fix it all by itself, at least at the moment.

It’s why our recommendation is to take a “process first” approach, so GPs make sure they have the right process infrastructure in place, the right organisational structure to manage the function, and then apply the appropriate tools to drive and sustain efficiency. It can improve almost any process, but there has to be a base level of competence to those procedures at the start.

Where should GPs begin to automate? How do they evaluate where automation can have the biggest impact?

“Where do we start?” is probably the most frequently asked question that we get. And I think GPs need to look at the process attributes of particular functions, say accounts receivable. Maybe there’s a big AR department and the outcomes aren’t great, with DSOs being unnecessarily extended. The next question is how easily such systems can be accessed via a technology platform.

Basically, it’s about deciding if it’s worth doing, or even feasible to do. And we find with accounting, HR and some sales functions, there are a number of places to leverage these solutions to great effect. Say, within the sales department, we’ve seen RPA reduce the close cycle from 30 days down to only 18.

What kind of upfront investment in time and money should GPs expect to make in these kinds of RPA solutions?

Most of our clients will have glaring places where RPA solutions make sense, say the finance or HR function, and we’ll get to work quickly. The complexity of the job will always influence the timing and cost, but in general they can get started on the low end, inside 45 days and for between $25,000 and $50,000.

Our process is a four-step approach. First, there’s an initial assessment of process environment, of the quantitative and qualitative aspects of a process, and whether they should do an automation project. Second, we’ll do a quick pilot programme on one of those processes. Once that’s completed, we will move on to the other relevant processes and as we dig into that, we train the user group so they can move it forward themselves.

How do you quantify the ROI on these types of investments?

Reduction of headcount is the most visible and immediate return on investment. That can prompt some scepticism among current employees. If a portfolio company has a large accounts payable team and this automation solution might remove 50 percent to 75 percent of that staff, there can be pushback.

But headcount reductions aren’t the only ways these bots deliver value. They can free up the current staff to do higher level work, like strategic initiatives, which might be more interesting than those very repetitive manual processes. And those processes will continue to improve thanks to an AI component that will be looking for new efficiencies long after the last worker went home.

How should GPs vet service providers in this space? How important is it that a service provider has experience implementing RPA solutions within a given industry?

Industry experience can help understand the infrastructure, but with this technology, functional expertise is probably more crucial than market expertise. The work that these solutions are automating in accounting, HR, supply chain and sales are broad disciplines where the fundamental processes are very similar.

However, within those broad categories, the service provider should have specialised teams devoted to continually improving those broad categories. That’s pretty much table stakes at this point. If a service provider can layer specific industry experience atop that, even better. Of course, it’s also a matter of the service provider offering references and case studies to help the GP select the right fit for the firm.

Once a GP has begun using RPA solutions, how can they stay current to ensure they’re deploying the best possible tool? Or does the AI evolve all by itself?

Those machine learning bots can improve processes all by themselves, but there’s always the possibility that some other solution will prove a massive leap forward. There are some options to tap outside consultants that can provide some independent analysis and advice on whether an upgrade is warranted, but I do think GPs can do their own homework online.

The truth of the matter is today’s bots are quite powerful, and do plenty without looking for the next best thing. For now, RPA solutions are still cutting edge.