Using robotic process automation RPA to create intelligent
Using robotic process automation (RPA) to create intelligent predefined tools Cameron Hobbs Mentored by Dr. Claude L. Jenkins III Introduction Materials and Methods The C++ coding language and Automation Anywhere Community Edition software were used to create the automation software for the Army SSC. Automation Anywhere is the leading enterprise in RPA development (Le Clair, 2017). To accomplish the identified goal of correctly sorting tasks among organization subdivisions, the software looped through each title and task description provided by the Army SSC. From these titles and descriptions, the RPA software subdivided the tasks according to similarities between task descriptions and an organization’s mission. Involvement for each subdivision was given a qualitative value within the RASCI matrix (Leopold, 2018). Both the developed RPA and the Army SSC employee were given the same matrix of Army tasks to sort. The time it took for the software to complete the matrix of assignments was compared to that of an Army SSC employee completing the same task by hand. Infeasibility, Rework 1. Initiation • Vision of a new or revised automation Ready for Development Redesign and Rework Infeasibilit y Rework 2. Development • Detail requirements • Testing of the proposed system Results 3. Deployment and Implementation • Deploy RPA tools • Monitor Ready for Operations and Maintenance Infeasibility Rework 4. Operations and Maintenance • Ongoing Improvements Ready for Implementation Terminate and Decommission Figure 1: A flow chart describing the development of all RPA systems. This was used for the creation of the RPA for the Army SSC. The general method for creating an RPA software is seen in Figure 1. The success of the RPA created in this project was determined by a 2 -sample t-test comparing the mean run-time of the RPA to a human sorting the tasks. The results of each completed assignments were compared to ensure an error rate that was less than or equal to that of the Army SSC employee. Results Graph 1 displays the time spent completing the task assignment by the RPA (n = 10, M = 23. 02, SD = 4. 144) and the human user (n = 10, M = 119. 33, SD = 8. 283). Time (seconds) Robotic process automation (RPA) is the ability to computerize monotonous tasks completed by a human (Anagnoste, 2018). Theoretically, RPA software increases efficiency in both time and money spent, as well as decreases error rate. It also frees human employees for cognitive tasks as opposed to data management. RPA serves to organize and manipulate large sets of data for methods designed specifically for a certain organization. The robotic process automation software created will be implemented by the Army Shared Services Center (Army SSC) to automate the process of assigning tasks to Army programs according to the responsible, accountable, supporting, consulted, and informed (RASCI) matrix. The purpose of the research was to create an RPA system to aid the Army Shared Services Center. Historically, a list of tasks was hand sorted into subdivisions of the Army SSC. It took a large amount of time to sort these hundreds of tasks by hand. The expected result of the intelligent automation software created to complete this task was that it would be significantly faster than a human user completing the same task, while maintaining at least the same level of accuracy. A functional RPA could be implemented by the Army SSC to complete the identified task of correctly assigning projects to subdivisions of their organization. Materials and Methods 140 120 100 80 60 40 20 0 Automation level versus the time spent completing a task Graph 1: Individual value plot of time spent by the RPA software and the human user completing the same RASCI matrix of Army SSC tasks. RPA Human User The null hypothesis for the 2 -sample t-test was the time spent assigning tasks by both the user and the RPA would be equal. Each set of times produced a p-value greater than the alpha level of 0. 05 when the Anderson-Darling Normality test was run. Therefore the 2 -sample t-test could be run. The 2 -sample t-test compared the mean time spent completing the assignment matrices by both the RPA and the human user. The test produced a p-value of 0. 00001 indicating there was a significant difference in task completion time between the RPA and human sorting. Both the RPA and human user produced matrices with identical RASCI classifications without error. Conclusion The purpose of the research was to create an RPA software to computerize the process of assigning tasks to subdivisions of the Army SSC. A 2 -sample t-test on run-times showed that the RPA was significantly faster than a human user. Unfortunately due to COVID-19, an employee of the Army SSC was not available to test. The human user was instead trained on best practice. Further research includes comparing the RPA to a true employee to see if there any differences in the result. If success is maintained, the RPA software could be implemented into organizations such as the Army SSC. In addition, the same task could be written in a traditional coding language or in an Excel Macro. The RPA software produced for this project could then be compared to these two alternatives. A similar test for time could be run to determine the fastest method for updating a task assignment matrix. References Anagnoste, S. (2018). Robotic automation process – the operating system for the digital enterprise. Proceedings of the International Conference on Business Excellence, 12(1), 54– 69. doi: 10. 2478/picbe-2018 -0007 Le Clair, C. (2017). The Forrester Wave™: robotic process automation, Q 1 2017 the 12 providers that matter most and how they stack up. Retrieved from http: /www. bluvaultsolutions. com/wp-content/uploads/2017/11/Robotics. pdf Leopold, H. , Aa, H. V. D. , & Reijers, H. A. (2018). Identifying candidate tasks for robotic process automation in textual process descriptions. Enterprise, Business -Process and Information Systems Modeling Lecture Notes in Business Information Processing, 67– 81. doi: 10. 1007/978 -3 -319 -91704 -7_5
- Slides: 1