The average time to hire a federal employee in fiscal 2017 was just under 106 days. But every manager knows that getting employees through the door is only half the battle.
There’s the dreaded paperwork shuffle of endless documentation and coordination to get new hires up to speed and equipped with the right technology to do their jobs. For the HR and IT staff charged with orchestrating many of these manual tasks, they are left with little time to work on higher-impact projects.
“There’s a lot of frustration across every workforce, and ours, where people have [these] mundane, data entry … projects that they have to do, and it needs to be done,” said John Lockwood, robotic process automation (RPA) program manager at the Defense Logistics Agency (DLA). “And they do it, and then that comprises the large portion of their workload, or maybe too large of a portion. And it doesn’t allow them to do the other things that maybe they were hired to do.”
To tackle these pain points, DLA is moving forward with several RPA projects, or investments in software bots that automate repetitive, rule-based tasks, such as copying and pasting information from one system to another or reconciling financial accounts. In August, the Office of Management and Budget directed federal agencies to move toward (RPA) in an effort to reduce the burden of low-value work.
DLA partnered with software firm UiPath on the project and estimates that it could save $2 million a year in productivity by onboarding new hires faster, Lockwood said. There are also time savings to be had once employees are freed up to do projects that they’ve had to put off. And by next summer, the agency is expected to automate about 50 processes using RPA.
Here’s How the Bot Works
To achieve those savings, here’s what the bots are enabling on the backend. When new employees come on board, DLA cybersecurity professionals no longer have to manually add employees to agency systems using a spreadsheet from HR. The work still needs to get done, but now the bot can do that task. Speeding this process up gives new employees quicker access to the software systems they need to do their jobs. The bots also enable cyber professionals to offload these tasks and focus on other critical work.
From a security standpoint, DLA is working to create credentials, such as a common access card, for what’s called unattended bots. These are bots that can operate independently without a human initiating work and actively monitoring that work. The reason is that when bots log into a system to perform work, the bots are viewed as humans from a systems perspective. Similar to any human, there are regulations around which systems bots can access and how much access they have.
Every click and every task that the bot completes is logged for auditing and security purposes. DLA is also building resilience in its RPA solution to ensure they can respond accordingly if there is a glitch or change. For example, if a bot is logging into an account to get information but that account is offline because a server went down, the software can be programmed to make additional attempts to access the system in an hour or so.
“An unattended bot can work 365 days a year, seven days a week, 24 hours a day, which is 4 1/2 times the hours that an individual would put in, ” Lockwood said. “A lot of times we want data running in the middle of the night. That unattended bot can do that with its credentials.”
He uses the term “digital worker” when explaining the technology to people across DLA. Using demos, he can effectively show them how a human would complete a particular process online and how a bot can do the same thing but faster.
Is Your Agency Considering RPA?
For agencies that are considering RPA, Lockwood shared some questions to consider:
- How much time does it take to complete the process?
- What percentage of the team’s time is spent completing the process?
- How many times is there an unusual case within that process that must be investigated or escalated to someone else?
- Are there any specific financial, regulatory or customer pain points within the process?
- Using percentages, how much of the process is 1) rules-based with structured input 2) rules-based with unstructured input 3) judgment-based, research and analysis.
- How complex is the process? For examples, does it follow very easy business rules to repetitively key information from one system to another, or does the process follow simple or complex business rules?
For more information about RPA, be sure to check out these GovLoop posts.