AI & Data Exchange 2026: Casey Mulligan on Driving Continual Processes Improvements

Why It Matters

By automating data‑intensive rule impact analysis, Sextant speeds small‑business regulatory reviews and sets a precedent for AI‑enabled efficiency in federal agencies.

Key Takeaways

• Sextant rebuilt legacy STATA tool in HTML/JavaScript via AI
• Tool cuts weeks‑long reporting process to minutes for staff
• AI‑generated prompts provide 0‑100 accuracy scores with source links
• Inter‑agency communications improve as AI reduces factual errors
• SBA’s AI rollout faces months‑long approval before public release

Pulse Analysis

The Small Business Administration’s Office of Advocacy is turning a bureaucratic bottleneck into a showcase for government‑wide AI adoption. Historically, the office relied on a single analyst to run a custom STATA program that evaluated the impact of new federal rules on small businesses under the Regulatory Flexibility Act. With more than 12,000 Biden‑era regulations pending, the manual process created a costly backlog. By prompting a large language model to rewrite the application in HTML and JavaScript, Mulligan’s team delivered a functional web interface—named Sextant—in a single afternoon, dramatically expanding access across the agency.

Beyond speed, Sextant embeds a sophisticated prompt suite that assigns an accuracy score to each generated sentence and automatically cites source documents. This built‑in fact‑checking mitigates the risk of “hallucinations” common to generative AI, ensuring that staff can confidently reference rules, policy statements, and agency guidance. The result is higher‑quality regulatory impact analyses, fewer back‑and‑forth clarifications with partner agencies, and a smoother audit trail. Early adopters within the SBA report that the tool reduces manual data wrangling and code‑syntax errors, freeing analysts to focus on higher‑value judgment work.

The broader implication for the federal landscape is significant. As agencies grapple with mounting data‑intensive mandates, AI‑driven platforms like Sextant could become standard infrastructure, accelerating compliance reviews and improving transparency for small businesses. However, the months‑long approval process to place the tool on a public website highlights the tension between rapid technological innovation and entrenched procurement and security protocols. Mulligan’s experience suggests that agencies willing to experiment with AI and streamline internal approvals may gain a competitive edge in delivering timely, data‑rich insights to both regulators and the businesses they serve.