The Department of the Air Force (DAF) has undergone significant changes recently to automate its procurement processes and address risk management issues, a DAF official shared on Wednesday.
Kori McNabb, a senior procurement analyst at the DAF, spoke about the successes and challenges the department is having in adopting AI throughout its processes at ACT-IAC’s “Emerging Technology Demo Day.”
McNabb shared that the DAF has used NIPRGPT – an Air Force Research Laboratory (AFRL) generative AI experiment – across its operations in the Space Force and Air Force.
“As of right now, today, every Department of the Air Force employee can access that on their computer without an individual cost of that unit,” McNabb said.
The DAF has used many of its generative AI tools to conduct risk assessments in procurement processes. McNabb said AI has allowed the DAF to reduce its risk assessments to occur between two and three hours, rather than days.
“It was a huge time saving and – more importantly – the questions that we were getting out of it were actually more robust than just going back and forth with the employees in the room just trying to crank questions out on your own,” McNabb said.
She also said a large portion of AI integration is getting agency employees comfortable with implementing it into their workflows. McNabb suggested employees use AI to complete the weekly “five bullets” directive where Department of Defense (DoD) civilian employees must list five things they did in the previous week on a weekly basis.
McNabb said one area for growth in automation and the use of AI is in reporting acquisitions to Congress, as required by law. She said using AI and automating the contract reporting process can increase the efficiency of acquisition reporting.
Additionally, McNabb highlighted data literacy inefficiencies throughout the DAF. She said better literacy will allow for improved decision-making for leadership throughout the agency.
“We’re trying to find solutions that can pull that data into one thing, and it helps our senior leaders make quick decisions based off what they’re seeing real-time versus information that’s maybe a month or two months in the backlog,” McNabb said.
