Making the most of your budget authority: AI-powered funds management

Federal agencies often struggle with managing budget authority and open obligations, leaving money on the table that could be repositioned for unfunded requirements or other mission-critical needs. Additionally, the review and oversight of unliquidated obligations (ULOs) and unmatched disbursements (UMDs) can complicate a financial statement audit and contribute to material weaknesses for many federal agencies.

The resulting ineffective funds management leads to reduced obligation authority and the return of unused funds to the US Department of the Treasury. The stakes are high: On average, federal agencies collectively return approximately $24 billion worth of budgeted funds to the Treasury each year. Returning undisbursed funds can negatively impact an agency’s future because Congress is typically unwilling to maintain or increase funding to agencies that regularly do not spend their entire funding allocation.

Pain points emerge from the processes and tools most federal agencies use to manage ULOs. One team extracts data from multiple systems (an ERP or financial systems and procurement systems), blends the data, and creates pivot tables. The team then emails the data to each division. Each division further distributes the records to its stakeholders for review. The division gathers all results of the review via email and sends them to headquarters for final aggregation. Follow-ups and reconciliations are rarely performed, or involve significant effort to ensure accuracy and completeness. Because agencies can have thousands or millions of ULOs and UMDs at a time, this approach can drain resources tremendously and divert staff from value-adding tasks.

AI provides an opportunity to better manage the overall budget process and better understand issues of cost estimation, budget formulation, and program management—all of which contribute to avoiding funds being returned to the Treasury.

The ideal potential applications for AI are where significant data volume complicates understanding or the drawing of conclusions. In such cases, organizations can add value by using AI to more rapidly analyze, classify, or interpret data compared to traditional manual approaches. In the context of budget authority, this could mean analyzing transactions’ risk of going unused before the funds expire. Then, management can focus on reviewing higher risk obligations, determine whether the funds are still needed, and de-obligate if necessary. Additionally, AI can be used to create and recommend budget authority scenarios, improving formulation and allotments.

AI-powered technology alone cannot address the challenges of obligations management and budget control. However, it gives humans a jump start on correcting recurring errors and improving overall fund management. UMD matching, automated high-risk ULO notifications, and dashboards to synthesize obligations and disbursements can be handled by AI. Then, staff can focus on finding and correcting chronic problems at the source. This can prevent future UMDs and reduce the total number of ULOs canceled each year.

Without the focus brought by AI, this outcome may be too difficult to realize for human experts. Agencies that balance the application of technology and human expertise may be rewarded with the optimal use of funding authority and greater visibility and control over funds.

Endnotes

¹US Government Accountability Office, “Federal budget: A few agencies and program-specific factors explain most unused funds,” May 25, 2021.

Get in touch

	Robert Gramss Advisory Principal Deloitte & Touche LLP +1 571 882 5240			Daniel Shorstein VP, Product Management Deloitte & Touche LLP +1 703 253 2710
	Haideh Chubin Advisory Managing Director Deloitte & Touche LLP +1 571 814 7140			Bryan Kaplan Product Manager Deloitte & Touche LLP +1 571 227 8583