Case Study

Accelerating government software delivery with AI-augmented engineering

How Mathematica’s software development team used agentic AI to deliver new capabilities faster for SAMHSA’s SPARS platform

Client
Substance Abuse and Mental Health Services Administration

Agentic AI enabled Mathematica’s software engineering team to deliver new SPARS features faster while maintaining the quality and reliability required for a mission-critical government system.

30-40%
Increase in sprint productivity
Client Need

Follow an ambitious roadmap without compromising quality

The Substance Abuse and Mental Health Services Administration (SAMHSA) needed new functions added to its Performance Accountability and Reporting System (SPARS) on an aggressive schedule. SPARS is SAMHSA's enterprise platform for collecting, managing, and reporting performance data from behavioral health grant programs, making it a critical tool for program oversight and accountability. Success depended on developing and deploying the new features quickly while maintaining the same levels of quality and reliability expected of a production system.

Our Approach

Integrating agentic AI into our software development

Mathematica modernized its software engineering workflow by incorporating OpenAI Codex into day-to-day development. By supplementing engineering expertise with agentic coding assets, AI helped the team move more efficiently from development through troubleshooting.

Our developers used Codex to generate code, investigate production issues, and support software integration. Staff reviewed, tested, and validated every AI-generated change before deployment, ensuring that human judgment, security, and quality control remained central to every production release.

Key Outcomes

Greater productivity and faster issue resolution

AI-augmented development increased our delivery capacity while helping us maintain the engineering standards required for a government production system.

  • Sprint productivity increased by 30 to 40 percent. Mathematica completed substantially more work in each successive sprint, helping the team keep pace with the project’s development roadmap.
  • We resolved production issues more quickly. After a successful data warehouse deployment, the team discovered a production issue that kept a new data download feature from working properly. Codex identified the problem as a configuration issue, enabling Mathematica developers to resolve it and release the feature to production within about two hours. Without Codex, the investigation would likely have taken much longer and delayed the release.
  • The client saw the results. Faster delivery of new functionality, combined with consistent engineering quality, helped the team meet demanding delivery expectations. By accelerating delivery of key capabilities such as SPARS data downloads, the team helped ensure that SAMHSA staff and grantees had timely access to the data needed to monitor program performance and inform decision-making.

Our takeaway

The SPARS project shows how agentic AI can help experienced software engineering teams work more efficiently when combined with disciplined engineering practices, a proven quality management system, and human oversight. The result was greater productivity, faster resolution of production issues, and timely delivery of new capabilities for a critical government system.

Partners In Progress

Melissa Azur

Melissa Azur

Principal Researcher

View Bio Page
Erika Krzton-Presson

Erika Krzton-Presson

Associate Director, Solution Architecture

View Bio Page

Efficiency Meets Impact.
That's Progress Together.

To solve their most pressing challenges, organizations turn to Mathematica for deeply integrated expertise. We bring together subject matter and policy experts, data scientists, methodologists, and technologists who work across topics and sectors to help our partners design, improve, and scale evidence-based solutions.

Work With Us