When it comes to process automation, new technology solutions are evolving faster than the vocabulary we have to describe it. The challenge for IT, finance and other business stakeholders is in sorting through a rapidly changing landscape of vendors and technologies to find practical solutions to real world problems with a reasonable return on investment.
When it comes right down to it, however, there are plenty of problems that can be solved today by software bots in a way that relieves human process owners from dull, routine work and relieves departments like finance from hiring staff members just to sort files and enter data.
Twenty years ago, robotic process automation (RPA) emerged as the next step in the evolution of what was once termed macros. They were simply screen-scraping tools that could replicate keystrokes and mouse movement, eliminating repetitive data entry for the person behind the keyboard.
In the IT world, much of this kind of work was considered ‘swivel chair’ work, copying data from one system to another, or moving a file to the right location. For high-volume data or document-dependent transactions, offloading this kind of work has great value. Three years ago, RPA solutions that could be implemented easily to solve these kinds of process bottlenecks were a big deal.
They still are-but now, the ability to tackle more complex tasks in the context of a business process is where the real excitement is. When it comes to back office processes like AP invoice processing or AR order processing, but there are all sorts of process gaps that require human intervention using simple bot.
Some call it cognitive RPA-others refer to it as intelligent process automation (IPA). Take your pick-with these technologies continuing to evolve, chances are that a new term or technology will come along to replace it just when one name sticks.
What’s more important that a name, however, is what cognitive RPA or IPA solutions do.
This is where artificial intelligence (AI) and machine learning (ML) enter the picture. While we generally think of AI as generalized intelligence that can adapt to a variety of problem sets (think IBM Watson playing Jeopardy), most practical AI applications today are specialized to handle a set of related tasks. Relying on machine learning, these systems not only imitate the actions of human process owners behind a mouse and keyboard. They can learn from those interactions, and they can identify and then predict patterns within data sets, or compare/contrast data sets to flag errors and correct them.
Where this gets interesting from a business standpoint, is when you look at the volume and velocity of business documents generated when companies purchase goods and services from vendors/suppliers, or invoice their customers and deliver goods/services. Historically, these operations have been relatively costly and inefficient. And in the worst of cases, they can slow down a businesses’ ability to grow to meet customer demands and tap into market growth.
Fast forward to right now, and you see companies putting intelligent process automation solutions in place to take all the kinks out of invoice processes that used to result in clogged email inboxes and mailrooms, requiring employees to spend the better part of a day behind a keyboard re-entering information that someone else entered on the front end of the process.
Today, intelligent process automation solutions like Artsyl’s docAlpha can leverage intelligent data extraction to identify and sort an invoice from a quote from an order, read the data off of those documents and then validate the information against ERP and accounting systems. It can match quotes, orders and invoices to fully automate PO invoice processing. It can route non-PO invoices to the right approvers and put the right coding information at their fingertips.
But what happens when the system finds and error? That’s where our human process owners still have a role to play. Intelligent process automation allows human operators to intervene to correct an error—in a way that minimizes the number of times a human has to do so, because the software adopts the new instructions into its algorithms.
That means no custom coding or IT intervention. Which further lowers cost and complexity.
To learn more, visit the Artsyl Resource Page