North Carolina cuts Helene loan admin costs 90% with AI

North Carolina's State Treasurer has deployed artificial intelligence to dramatically reduce the overhead of managing disaster recovery loans, allowing the full appropriation of nearly $23 million to reach communities still rebuilding from Hurricane Helene.

State Treasurer Brad Briner announced Tuesday that his department eliminated administrative costs entirely for the latest round of interest-free loans to local governments affected by the deadly September 2024 storm. The department had initially anticipated needing additional staff or an outside accounting firm to manage the program.

How AI reduced overhead to zero

State law permits up to 2% of appropriated funds to cover administrative expenses for the loan program. For this distribution—$20 million newly appropriated in 2026 plus $2.7 million remaining from earlier rounds—Briner's office is taking nothing for administration.

The department is using AI and automation to track 119 different borrowers, each with distinct loan agreements, repayment schedules, and deadlines. The technology identifies when borrowers need contact and handles routine tracking tasks that would otherwise require manual spreadsheet management.

"We've observed that we can do this much better with AI than we ever thought we could without it," Briner said in an interview with WRAL Wednesday. He characterized the approach as removing complexity and tedious work, freeing staff to communicate directly with local governments rather than wrestling with data management.

Demand far exceeds available funding

The efficiency gains come as demand for the program significantly outpaces resources. Treasury received more than $100 million in loan requests for the 2026 round, leaving roughly $80 million in requests from Western North Carolina communities unfunded.

The cash-flow loan program, created in 2025, provides local governments with bridge funding for rebuilding and cleanup while they await federal recovery dollars. All loans must be repaid without interest by June 30, 2030, after which the funds can be loaned to other communities.

Briner indicated his department stands ready to expand the program if the General Assembly appropriates additional money when lawmakers reconvene.

Why it matters

This deployment demonstrates how AI can directly increase the impact of public spending by eliminating overhead rather than replacing workers. In disaster recovery, where speed and resource allocation are critical, the 90% cost reduction means communities receive funds faster and in larger amounts. The model may inform how other state agencies approach program administration, particularly for time-sensitive initiatives where manual processes create bottlenecks. As state governments face pressure to do more with constrained budgets, North Carolina's approach offers a concrete case study in operational efficiency through automation.

Broader AI adoption in state government

The Helene loan program represents one application within a wider AI expansion across North Carolina's state operations. Briner previously partnered with OpenAI on a pilot examining how ChatGPT could improve government functions, including analyzing public financial data and identifying process improvements.

After several months establishing safeguards, policies, and contractual terms, the State Treasurer's office rolled out OpenAI tools across the department earlier this year. Multiple divisions now use AI, with specialized tools being added for specific teams.

"We have to get more efficient as government," Briner said, noting frustration both inside and outside government with operational speed.

These details were first reported by WRAL, with reporting by Caroline Yaffa.