At MITRE—a private, not-for-profit company that provides engineering and technical guidance for U.S. government agencies, including defense, healthcare, and cybersecurity—AI expertise is not in short supply. However, finding the right experts within the sprawling organization to inform projects has been a major challenge. With more than 9,000 employees, 2,000 live projects in any given year, and headquarters in both Virginia and Massachusetts, the organization is too big for even an experienced project manager to navigate without help.   

According to Dr. Cedric Sims, a senior vice president who heads MITRE’s enterprise innovation and integration function, an in-house AI-powered application known as 6DOS (short for “six degrees of separation”) is the solution. This capability helps teams find connections between employees based on the information they share via internal email, messaging, and meeting programs. But 6DOS is just one way the federally funded nonprofit is deploying AI to tap into its wealth of collective knowledge.

Among the more active workflow discussions are those had by project managers. Dr. Sims points out that these employees are responsible for reporting progress updates to project “sponsors,” or clients within government. One of MITRE’s lead project managers is currently working on a GenAI prompt that aggregates inputs from the prior week’s project management reports, transcripts of meetings that have occurred during the week, calendar data, and the like to “create a project management report for the current period of time that automatically provides the accomplishments against to-dos and works in progress and then authors the next set of anticipated work products,” says Dr. Sims.

Thanks in part to a community improving this prompting initiative, project managers can now generate reports for sponsors in a matter of minutes rather than hours. But for all the efficiencies that AI can bring, MITRE’s more senior, longer-tenured practitioners want to ensure that this automation does not compromise the intimacy that the company has with its sponsors. “We don’t want the advancement of these capabilities to cause folks to not do the necessary human-in-the-loop validation,” says Dr. Sims. “You need to have a discriminating eye and confirm the information and details. So, it’s a job aid, not a replacement.”

The nonprofit also places a premium on balancing risk and reward by giving a lot of responsibilities to employees while still maintaining visibility across activities. That visibility extends to MITRE’s CEO and the board of trustees, who oversee the implementation of AI and its implications to MITRE’s future. Dr. Sims credits a sophisticated workforce for making it easier to strike that balance and a clearly articulated AI strategy informed by senior managers that ensures “investments regarding AI are meeting the collective needs of the organization.” A recently established AI council, a direct outgrowth of the organization’s AI strategy document, is intended to establish greater understanding of sector needs and how well MITRE is meeting them.