Intelligent process automation solutions are quickly transitioning from pilot projects and initial test phases into real world applications that live up to their hype — particularly when it comes to automating back office business processes. From the standpoint of the trade press, this is a two-edged sword. One one hand, the hype and promised of “AI-powered” applications sound cool and interesting, warranting interest and readership. It also invites a healthy bit of wait-and-see. When the hype is translated into practical, real world applications, and those applications wind up being the kinds of traditional (and less than cutting edge) business processes that have been the subject of automation for decades, it takes the luster off the promise of the new.
That said, if you dig into the details and you care about process efficiency, customer satisfaction and all the things that lead to a more profitable, competitive and successful business, you have to admit — there are some pretty cool things going on in today’s intelligent back office.
For one thing, the work involved to automate a simple business process has lost a lot of the overhead it used to have. The work we’re talking about here is the effort just to get a system up and running in a way that automates manual tasks in a way that is sustainable. In the past, rules-driven workflows required an IT person to configure and maintain, and most likely a developer to customize integrations or rules-based workflow steps that weren’t common to a business process.
Robotic Process Automation (RPA) changed all that, for very simple tasks, at least. They could allow you to automate a mouse click or a keystroke in a way that didn’t require a developer to work with an API toolkit to do so within an application. Swapping RPA for API isn’t just a matter of trading one acronym for another. The power is that companies can use RPA to overcome the complexity of automation by using tools that can record a human executing a process and translate that into an automated process flow that a process owner can further tweak and modify to automate a task in a way that can work across applications. This approach can overcome all kinds of issues, including not having to worry about system integration to get data from one system to another. Rather than integrating, you allow the bot to handle the kind of keystrokes and mouse clicks your process owners used to have to do manually, over and over again.
Then, things got even more interesting. Companies demanded that these easy-to-use automation tools do more. They needed them to be able to go beyond simple, routine tasks to handle more complex process steps like those found in high volume operations like accounts payable invoice processes and customer service/sales order handling.
After all, if RPA tools and record, analyze and automate keystrokes from a human operator, couldn’t it learn from that human operator and adapt on the fly to changes in operations to address process exceptions on the fly?
Answer: yes, absolutely, with the application of machine learning. Combine machine learning and RPA and you get intelligent process automation.
Intelligent process automation allows companies to cost effectively deal with business processes that depend on unstructured content (emails, business documents, etc. were data elements aren’t discretely organized and structured in the way a form might be), workflows with a greater degree of variability and systems that may require a higher degree of integration and interactivity that simply translating mouse clicks and key strokes.
These solutions reduce IT and implementation overhead, radically improve system flexibility and truly put the power in the hands of the user. The compelling ROI scenarios for automating AP or Sales order processes applied to traditional rules-based workflow solutions, but today, with most of the complexity taken out of the equation, it’s become nearly impossible to look at manual back office processes as a cost of doing business. We’ve reached the point where the cost and consequence of the status quo could be the failure to compete in a more agile, automated marketplace.
Intelligent Process Automation initiatives have gone from discovery projects to explore what’s possible to a focus on real world applications that target the most painful, costly and less-than-exciting business process. That’s good news for business, even if it doesn’t always lead to the most compelling headlines.
For business looking for ways to take advantage of new automation approaches, our recommendation is don’t start by sizing up the technology. Start by looking inside your own organization for sources of waste — both in terms of cost and human effort.
Chances are, there are plenty to choose from. If you can define them in terms of current costs and business impact and then rank them by benefit, that’s a great start.
If that’s too much work, or you lack the data to accurately or confidently quantify cost/benefit, then look to your peers, whether by industry or company size. What processes have they successfully automated and proven out an ROI?
Chances are, you’ll find some common themes. That’s because IPA tools have established a track record for solving for common back office business process pains. Accounts payable vendor invoice processing and sales order processes are common culprits and often represent low hanging fruit when it comes to process automation.
Not that we’re trying to lead the witness here. But, companies often find that the investment they make so automate once process, when they invest in a platform that is build to automate a wide range of processes, allows them to re-apply much of their technology investment, and their experience with IPA, to other business processes downstream.
Once you agree on what to automate, be mindful about the benchmarks you establish for success, and how you measure them. Hyper-inflated expectations about “AI” can lead people to think that all you need is a massive pile of data and an AI tool, and the system will do the rest.
That’s a massive oversimplification, of course, but don’t let any vendor or consultant convince you that there isn’t work involved in getting an IPA implementation right. Most of that work is not technical — it’s more around understanding your business processes well enough to map out the right solution and put it into practice.
The good news about IPA is that the cost of getting it wrong isn’t what it used to be. Because it is easier to configure and adapt IPA processes than it was to build out rules-based workflow systems, you can get them up and running, test them and modify them more easily.
But nothing can make up for setting the wrong expectations with the stakeholders who approved and sponsored your initiative. You still need to set clear expectations and do the work of defining your processes up front.
We can all apply the lessons learned from designing and deploying any business system, including ERPs, ECMs and CRMs to intelligent process automation. Define the data you need to inform your process, define the metrics your want to monitor and manage up front. Don’t assume your already have what you need in any system your IPA tool may rely on to validate a vendor or customer record. Don’t assume that the KPI you need to track is already baked into your AP automation solution or your sales order automation application.
All too often, process automation solutions go too far downstream before the process owners are involved. IPA truly puts those process owners in the driver’s seat. They should be central to the project, with support and guidance from your business analysts, IT and development staff.
This is equally true up front when deciding what use cases to automate, as well as throughout the process of designing, testing and deploying the solution. While this best practice has always held true, it is more true and more applicable today than ever before.
When it comes to vendor invoices or sales orders, you’d might assume that existing processes are well-defined processes. We all know what happens when we assume.
Despite broad commonality across industries and business size, these processes DO vary from organization to organization, and the devil is in the details. It also depends on which process owner or stakeholder in the process you talk to.
Even so, understand that truly understanding the process may be iterative. You may have to try things out before you really get to understand where the underlying bottlenecks or process issues are.
The good news is that IPA solutions are far more friendly to that kind of iterative approach than old school rules-based systems
To learn more about how to evolve your processes from manual to automated to intelligent, contact Artsyl Technologies and request a demonstration of the ActionSuite of intelligent process automation applications for accounts payable, accounts receivable, claims management and more.