Artificial intelligence is proving to be an integral technology for improving federal procurement and reducing workforce burdens.
“Procurement really is a critical enabler for any organization to meet its mission,” said Shanna Webbers, Internal Revenue Service’s chief procurement officer, during GovernmentCIO Media & Research’s AI Gov: Data Insights virtual event.
Historically, government’s culture has emphasized compliance and manual processes that have prevented innovation, making it difficult to meet increasing workloads as technology advances. The IRS is one agency leveraging AI to improve these workflows and expedite processes across its procurement arm, Webbers said.
Using AI in the acquisition process should fall under three goals: automate, optimize and enhance.
“I like to refer to it as a shift from tactical to strategic, allowing your workforce to do more with less, and freeing up time for more strategic and complex tasks,” said Mike Cook, general manager at Icertis, said in outlining those goals. “With AI, intelligent automation can be a big enabler for agencies to better deliver on their mission with that laser focus on outcomes.”
Over the past few years, the Department of Homeland Security has improved its Contractor Performance Assessment Reporting System (CPARS) system, which agencies use in their contract award decisions to measure quality and timely reporting of contractors’ performance.
Traditionally, contracting officers have heavily relied on written proposals to select the team or the best vendor on a contract. With over 1 million records for tens of thousands of vendors in CPARS, accessing that data can become difficult to manage, said Scott Simpson, digital transformation lead of DHS’ Procurement Innovation Lab. Simpson explained how the agency is applying AI tools to the system to allow contracting officials to more easily search for information.
“We don’t want the system to make decisions on behalf of the contracting officer … we are augmenting their decision-making process by helping use AI, like natural language processing, to surface the most relevant reports and evaluations to the top,” Simpson said.
Simpson said the agency has incorporated nine other agencies for access to the tool.
Webber noted that the IRS leverages the same process. The one-time build method will bolster consistency and ensure that agencies have access to the best products on the market that meet specific requirements.
“Allowing us to have consistency across the federal government frees up others to create new tools that then can also be leveraged. That multiplying capability is what's really going to be game changing for the acquisition community,” Webber said.
AI and predictive analytics would provide contracting officers with the ability to data-mine billions of contract transaction records to compare past experience, complexity and performance, Webber said.
IRS has also focused on workforce development, having partnered with academia and small businesses to create tailored data analytics and procurement curriculum for the contracting workforce. This training will create a “data literate” organization that can understand technologies’ output, assimilate information and drive better decision-making.
“It's really important that we recognize that not everybody will be the same, and we don't need the same skill sets, but we do need a baseline and a foundation of understanding of how to utilize and leverage outcomes from these very powerful tools that are providing us with information,” Webber said.
IRS integrated an “intelligent risk taking” approach, where teams identify and understand risks, then move forward by integrating new tools. With robotic processing automation, for example, Webber’s team was able to complete a year’s worth of work in three days.
Moving forward, Simpson projects that AI will change market research by plugging in a few key words, or machine read solicitations to provide new vendor and product insights “that take us into the 21st century of procurement,” he said.
As the government continues to adopt emerging technologies, Cook noted that there will be increased focus on ethical developments of AI and automation.
“When we're looking at fair and open competition, how do you continue to monitor and make sure that your algorithm is fair, impartial, transparent and explainable? From a security standpoint, we also have to make sure that it's secure and has that ongoing monitoring,” Cook said.
“By allowing these learning algorithms to strategically predict when these key events will occur is going to be huge for us,” Webber said.