Sheng Wang
Principal Data Scientist
View Bio PageCase Study
Accelerating regulatory analysis with generative AI
Smarter comparisons, faster insights: Using AI to streamline CMS’s annual IPPS rulemaking
Client
Automating complex rule comparisons with AI, while keeping experts in the lead, dramatically accelerated IPPS rule analysis without sacrificing rigor or accuracy.
94%
estimated reduction in annual costs (approx. $10,000)
Client Need
Navigating thousands of pages of complex regulatory text
Each year, the Centers for Medicare & Medicaid Services (CMS) issues the Inpatient Prospective Payment System (IPPS) rules, which govern how hospitals are reimbursed for inpatient care. To prepare each year’s updates, policy experts must compare the new rules with prior versions and identify every relevant change, a process that requires weeks of manual review, tracking, and cross-referencing.
But the volume and complexity of the text makes it difficult for experts to focus on higher-value policy analysis, creating a need for a faster, more reliable way to navigate large regulatory documents and ensure accurate rule comparisons.
Our Approach
An AI-powered tool to retrieve and compare rule language instantly
To streamline this process, Mathematica developed a chatbot powered by generative artificial intelligence (AI), hosting it on Amazon Web Services’ Bedrock service to ensure secure, compliant, and scalable cloud deployment. This solution enables Mathematica’s experts in CMS policy to explore IPPS rules using a chatbot.
The chatbot quickly retrieves relevant rule sections from the current and prior year, highlighting differences and streamlining staff verification of updates. A rigorous “AI + human-in-the-lead” process enables our experts to continue confirming the accuracy and context for every result.
Key Outcomes
Faster turnaround, better accuracy, and more time for expert analysis
The AI solution transformed the efficiency and precision of Mathematica’s IPPS rulemaking work for CMS:
- Drastically reduced turnaround time and boosted productivity: The response time for each comparison request dropped from 3 or 4 hours to less than 30 minutes—a major savings for an activity performed up to 10 times per year—estimated to reduce annual costs by about 94 percent (or $10,000). Policy experts spent far less time on manual review and could focus instead on higher-value analysis, interpretation, and policy insights.
- Improved accuracy and consistency: AI-assisted retrieval ensured that experts could capture and verify all relevant changes, reducing the risk of omissions or misinterpretations.
Beyond the IPPS rules, Mathematica experts are now using the chatbot to:
- Summarize CMS payment models and identify policy trends
- Explore new quality measures and emerging policy directions
- Enhance transparency and accelerate decision making across CMS program areas
Our Takeaway
This work shows how pairing generative AI with deep policy expertise can dramatically speed up complex regulatory analysis without sacrificing rigor. By automating time-intensive document comparison and embedding human review throughout the process, Mathematica helped CMS move faster, reduce costs, and improve consistency, while freeing experts to focus on interpretation, judgment, and policy insight. The result is not just more efficient rulemaking, but a scalable model for using AI to support smarter, more transparent decision making in health policy and beyond.
Partners In Progress
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