Common RPA Myths Debunked – Part 1

HPA News

RPA myths debunked

The RPA industry can be a place of buzzwords and a whole lot of hype. The marketing misinformation pushed in this industry creates a lot of confusion and leads to unrealistic expectations on what can and can’t be achieved with automation. In Part 1 of this series, we're taking a no-hype approach to unpack those common RPA myths and set your automation initiative up for success. Be sure to catch Part 2 here.

Myth #1: Robots can "go rogue" and make decisions on their own

When most people think about robots, they envision some Hollywood version like I, Robot that takes on a mind of its own and begins destroying everything in sight. While there are robots that can learn and make decisions on their own, the reality is that robots are pretty boring. The majority of RPA robots are "unattended" robots, meaning they follow a specific set of instructions and execute those instructions start to finish with no external inputs from a human. These types of robots will generally work off specific, defined data points and interact with fields in the business applications. Robots are configured to only follow the explicit actions and decisions dictated by the business process, and nothing more. The small percentage of robots that can make decisions on their own require extensive, continued training and must be monitored by humans to ensure their outcomes are correct. Advances are being made here, of course, but it is incredibly difficult to replicate human decision-making with consistency and accuracy.

While we can't speak to other people's robots, HPA’s robots use positive and negative validation at every step of the process to ensure that the robot is seeing what it expects to see, and to identify and alert on what it doesn’t expect to see so that unexpected outcomes don’t occur. If the robot encounters anything outside of what it is "trained" to do, it will immediately stop and back out of the transaction, noting what it encountered that was out of scope and alerting our team to investigate. You want this level of security built in to your robots because there tends to be a waterfall effect that happens when one process in a series of processes is run incorrectly. With validation safeguards in place, it can shield you from the fallout of a missing process step in the robot instructions, an incorrect data set, or changes to the underlying applications that the robot has not been instructed to account for.

Myth #2: Robots can't read or consume data

While robots don't have eyes, they are able to read and consume any type of data that is digitized and structured. What I mean by 'digitized and structured' is data that is formatted or labeled, such as a database table or an Excel spreadsheet, or data that can be extracted from documents and formatted for robot consumption. Robots are able to read and interact with fields and data points within business applications just as a human would. Robots are also able to interact with external systems, applications, and APIs. They can navigate to an external website or to Microsoft Office Suite to pick up data for use in a process. Plus, unlike their human counterparts, robots are able to consume all data required to run a job on the front end, rather than gathering data as they process the item. For example, when comparing data from two different systems, a human would need to flip back and forth between the two. Terribly inefficient and prone to error. Robots are able to consume all the data they need to execute the process in one pass, then compare or replicate that data in the secondary application with one fell swoop.

Myth #3: Automating a process will improve it

The term "garbage in, garbage out" applies here. Inefficient processes breed inefficient robots. Robots runs on set business rules. If those business rules are inconsistent or constantly influx, your robots will be running on incorrect, incomplete or outdated rules. This will give you incorrect, incomplete or outdated results. Likewise, if the applications your robots use are unstable, the robot will also be unstable. Both of these scenarios are the top causes of robot instability and change management. Automation can actually help expose inconsistencies in your process, but those inconsistencies will need to be addressed in the robot instructions in order to generate the results and ROI you are looking for. Solid business rules lead to great automation.

In Part 2 of this series, we will unpack four more RPA myths around robot reusability and common assumptions about how automation should be applied in your business.