I recently listened to a podcast in which Tom Romeo of MAXIMUS, which is a large contractor of the US Federal government talked about how his company has chosen RPA to automate many services provided by the government. It was striking to hear how far along in thought he was regarding RPA and the use of AI.
Much of the directive germinates from an executive order that was signed on 2/11/19 that essentially mandated that US Government use AI and move forward with their federal data strategy.
In the podcast he described the origins of RPA (i.e. scripting and macros) and contrasted it to where it is today. Implementing best of breed techniques as well as a host of tools to make it more powerful (i.e. Machine Learning, AI). RPA activities themselves are evolving and getting more robust. However, the tools that are available now outside of RPA are enhancing the platform to achieve greater results.
At CampTek Software we’re constantly looking for ways to improve automations and fill in the gaps that have plagued the space for years. While things are more reliable than ever, there are applications and decision points you can’t automate using typical means. RPA as a rule requires instruction and it isn’t very smart. RPA can’t make decisions on its own without the help of a human, use of AI, ML, NLP or even an API of some sort.
This doesn’t minimize the power of the tool at all. To the contrary, since most platforms like UiPath allow one to easily add custom activities that are either compiled assemblies or make calls to cloud based API’s. This can exponentially extend the capability of the tool. Not to mention the ability to write code using the IDE to spin up a method using referenced assemblies or your good old fashion ingenuity.
Having this power really opens up many possibilities. The most common use case we use AI and ML for is reading unstructured and varying formatted invoices. This is an age-old problem that many companies have done “well”, but I am of the mind that the typical approach doesn’t deliver the results that are truly satisfactory.
We use several techniques in this scenario that can simplify the implementation and maintenance of ever-changing invoices. Our developers aren’t spending countless hours trying to parse a string or figure out positionally how to pull out the invoice. Without getting into specifics I will say we have been very successful employing these techniques.
This is only one-use case, but intelligent automation is the way to get the best bang for the buck. I know AI, particularly ML and NLP are going to become more commonplace in every area of business. Using deep knowledge captured by humans in the areas that require expertise are going to explode. Think about the functional areas of Law, Accounting, Invoice processing etc. All are lines of work that generally require domain knowledge in addition to a set of common principles.
If these were put in a decision engine of some sort, the effects can be powerful. Essentially all the dependencies of a decision would handled by an external decision making resource versus programming every dependency within an automation.
This is truly an exciting time to be alive, things are going to get much easier for everybody. As always, it may create new challenges, but we will be ready for it!
Peter S. Camp is the CTO and Founder of CampTek Software. He has been developing RPA Applications for over 15 years. For further questions, discussion or inquiry about CampTek Software Services, contact firstname.lastname@example.org.