“That’s not right!”
“I thought everyone did it the same way!”
These are common reactions when employees discover their colleagues are entering data differently than they are – or that they are using it in different contexts. Exasperation is also common – especially when people realize they are basing reports and decisions on questionable information or that they will have to do work over. When staff in one organization lacks trust in another’s reports, they end up reaching out to trusted contacts to get the right information instead of using what is supplied to them. In the meantime, the company is busy generating, questioning or correcting bad data. Recent studies claim that up to 20 percent of corporate data is flawed. All told, data errors cost the U.S. economy trillions of dollars.
Humans are prone to error and it is well documented. Dr. David J. Smith has studied reliability and safety engineering since the 1970’s and written and updated numerous textbooks. According to his human error rate table in Reliability, Maintainability and Risk, a human fails five percent of his or her attempts at entering 10 digits into a field and reads information incorrectly one percent of the time. The human error table lists the error rates as “complicated non-routine task” (1 error in 10), “routine task with care needed” (1 error in 100), “routine simple task (1 error in 1000) and “simplest possible task” (1 error in 10,000). Multiplying error rates against data-related tasks can multiply costs quickly. Human operators become bored, distracted or interrupted during a work day. And, over time, on-the-job training or lack of formal training also takes its toll on data accuracy.
Robotic process automation (RPA) consistently and quickly completes repetitive tasks – the exact kinds of tasks a human is likely to mishandle. Robots can compare invoices to purchase order information to approve payment, including research to resolve errors. They can review claims to validate information, and even send a correspondence to the claimant to correct or provide additional information. These software bots use the same user interfaces that human operators use without making the errors humans are prone to make. Bots accomplish tasks without becoming bored or distracted when repeating the same task, and they do not suffer from one-off training or poor memory. Finally, bots do not read data incorrectly or make “fat finger errors.” If bots encounter a situation for which no business rules exist, the bot can log the issue and notify a human operator to handle it. Then the bot will press on with the next transaction.
RPA software offers the benefits of 100 percent accuracy, reduced time to complete and lower labor costs for menial tasks. Transaction activities using structured data are perfect candidates for RPA solutions. And not only do they eliminate errors, bots free up human operators to do work that requires knowledge worker skills – activities that can drive your strategic plan and generate growth.
ISG helps enterprises explore the widening RPA opportunity. Drop me a line to discuss whether RPA software is right for your organization.
About the author
Richard Baldwin is an experienced consultant with more than 30 years of experience in information technology, process automation, transformation and software development. His consulting experience includes work for firms such as Accenture, GenPact and Toshiba. His industry experience includes Department of Defense, energy and utility, retail, finance, and not-for-profit industries. He has held executive and senior technology positions for consulting and finance firms.