The Leader’s Role in Implementing Artificial Intelligence and Robotic Process Automation
Artificial intelligence and robotic process automation are being used to help federal agencies in their digital transformation strategies. As part of my detail with the IBM Center for The Business of Government, I met with several thought leaders across government about this topic, as well as practitioners across IBM. This helped me glean key insights on how leaders may incorporate robotic process automation (RPA) as a precursor to artificial intelligence (AI) in their workforce and workplace.
For example, I had the opportunity to hear directly from Margie Graves, the Deputy Chief Information Officer for the US, at a meeting of the National Academy of Public Administration this past July as she touched on ways to tackle implementation of AI. She also described that the best way to view AI is in three different phases: assisted AI, augmented AI, and autonomous AI. These phases can be ranked from least to most risk in terms of execution. I will focus on assisted AI which looks at repetitive tasks and it is where we commonly find RPA in use. RPA is closely related to automation technology an area I am familiar with due to my military service.
We all know there are some things only humans should carry out in the workplace. For example, tasks such as personnel counseling or an individual’s performance reports are vital to keeping an efficient and effective workforce. However, data entry into a system or repetitive tasks are optimal choices for utilizing RPA. Currently, those manual tasks done to supplement interoperability between systems or reports are exactly the first processes for the use of RPA to safely execute with minimal risk to the organization.
To ensure smooth implementation of RPA in the workplace and workforce, leaders need to act in three areas: process, privacy and policy. These three focus areas are keys to success for leaders to swiftly implement the gains in efficiencies from RPA and avoid disruption in achieving organizational goals.
Restructure Process to Optimize the Use of RPA. Process can be thought of as a high-level view of your way of doing things, whether it’s written in documentation or something culturally occurring within your organization. Process drives how things are done and is taught formally or informally. Most of the time it’s the on the job experience leading to detailed procedures learned when you begin working in a new capacity.
Process presents the least amount of risk to organizations as it is easily grasped and trained. Using RPA to replace accounting and repetitive tasks has been found to provide the opportunity to use this time elsewhere. For example, U.S. Customs and Border Protection was determining how to access and index 30 terabytes (TB) of its archived email. The emails were stored in a file format that required manual conversion and it took 2 months to convert each terabyte of data. Using RPA, however, unattended bots assisted by AI were able to move, convert, and index these email files much more efficiently than doing it manually. This resulted in the conversion of 30 TB of email files into a modernized platform, thus reducing manual processing time from 22 months to 3.5 months. This change in process allowed agency leaders to redeploy 24 full time employees to focus on high value, mission-critical tasks.
Embed Privacy Strategies to Ensure Employee Trust. Data privacy is a key consideration when implementing RPA within organizations. It’s clear that information gleaned from data is only as good as the source providing the data. At times this data is proprietary or sensitive in nature creating privacy concerns when it is desired to share information across organizations. A common way to access and validate the identity of those accessing the data is necessary.
For example, a known RPA vendor has been using robots for the past two years to validate employee credentials. In working with the Defense Logistics Agency, it used common access card (CAC)-enabled sites to develop a solution to allow government agencies to use robots in their process. Likewise, the personal identity verification (PIV) card permits robots access to the NASA Shared Services Center systems. Using this approach provides trust in the security of the data and enables the RPA to function while maintaining data privacy. Additionally, leaders must be engaged to ensure leakage, or the unwanted sharing of personal information is not occurring.
Policy Changes Must Accompany Culture Changes. Policy is difficult to implement and enforce with RPA because the technology is relatively new. The Office of Management of Budget (OMB) is leading the charge in implementing policy to address the February 2019 Executive Order on Maintaining American Leadership in Artificial Intelligence.
In its follow up guidance, OMB directed federal agencies to use RPA to develop and implement strategies for shifting resources to high value activities:
“Agencies should develop and implement reforms to eliminate unnecessary or obsolete compliance requirements and reduce the cost of mission-support operations. Reforms may include streamlining or eliminating unnecessary reporting requirements, consolidating processes and functions across offices, using shared service solutions or technologies, eliminating agency specific guidance or policies that preclude using shared services, and introducing new technologies, such as robotics process automation (RPA), to reduce repetitive administrative tasks, and other process-reform initiatives.”
Leaders implementing policy changes must also focus on the culture of their organization. An article written by Army War College students does an outstanding job highlighting the strides made thus far in private industry integration of AI into their culture and how leaders can leverage these lessons within their own organizations. For example, “AI can help unlock intraoperative procedures with video data.” Despite being counter to the culture of medicine, recording surgeries can help detect deviations during surgeries and prevent complications. Gathering such data requires a cultural shift in thought for surgeons reluctant to have their surgeries recorded. The knowledge gained and shared across the organization would address the need and the why behind creating policy and adjusting organizational philosophy.
In closing, the use of RPA and AI is vital to shifting our workforce to more high value tasks addressing the key needs of an organization. Using current examples found in industry and government provides leaders with considerations to mitigate risk in the areas of process, privacy and policy. Proper training and workforce education will go a long way to driving implementation and allowing naysayers to embrace this digital transformation.
Major Tracy Tawiah, U.S. Army, is an active duty operations research/systems analyst currently serving as a Research Fellow in the Army’s one-year-long Training with Industry Program at IBM. Her research fellowship is intended to develop her technical expertise and analytic leadership so that she can apply industry best practices in her future Army assignments. This blog post was written during her time at the IBM Center.