Providence Health, one of the largest health systems in the U.S., launched an ambitious initiative to improve access to care and reduce operational strain through AI-powered messaging. With rising administrative costs and increasing demand for self-service, Providence recognized that better patient engagement could also drive stronger revenue performance. Rasa became the platform to support that vision.
Key Takeaways
- Providence chose Rasa to help scale its conversational AI agent Grace across multiple service lines and patient journeys
- Using Rasa's CALM (Conversational AI with Language Models) architecture, Grace evolved into a more sophisticated AI assistant that can manage longer, more complex conversations and support even greater levels of automation
- The agent now handles 160K+ unique monthly user conversations and 106K+ monthly task completions
- Improvements includes 59% goal completion rate per conversation and 28% repeat engagement rate
Healthcare systems often face bottlenecks in access, leading to missed appointments, delayed care, and staff burnout. Providence saw the opportunity to address these issues head-on by streamlining how patients get information, manage appointments, and engage with their care journey.
Their goals focused on reducing manual touchpoints in common patient workflows, supporting care teams with fewer interruptions from administrative questions, and improving revenue performance through more successful appointment scheduling and follow-through.
Challenges
Providence’s Digital Innovation Group (DIG) developed Grace, a conversational AI agent that supports patients through secure, personalized messaging.
Providence built a dedicated cross-functional team to support Grace’s evolution, combining product, design, data science, and engineering. The team’s focus on patient-first workflows and measurable outcomes helped Grace deliver results across digital channels while supporting Providence’s long-term engagement strategy.
After evaluating multiple platforms, the team selected Rasa to help scale Grace across multiple service lines and patient journeys. They prioritized speed to accelerate assistant development across teams, flexibility to enable participation from engineering, UX, and analytics, and security to support compliance standards and on-prem infrastructure. Rasa enabled the team to build and iterate quickly while maintaining the standards required for safe, scalable deployments in healthcare environments.
Grace now assists thousands of patients each month with tasks like booking appointments, rescheduling visits, and navigating digital services. By supporting more proactive communication and reducing drop-off in care journeys, Providence has seen stronger engagement and improved follow-through on key revenue-driving interactions.
Highlights include 160K+ unique monthly user conversations, 106K+ monthly task completions, a 59% goal completion rate per conversation, and a 28% repeat engagement rate. Ongoing work with LLMs and transformer models has improved Grace’s ability to understand patient requests and route them effectively, helping more patients complete their tasks without escalating to human support.
Results
Providence is now expanding Grace’s capabilities using Rasa’s CALM (Conversational AI with Language Models) architecture. The goal is to evolve Grace into a more sophisticated AI assistant that can manage longer, more complex conversations and support even greater levels of automation. CALM separates conversation understanding from execution, allowing LLMs to interpret requests while keeping control logic structured and versioned. This gives the Providence team the flexibility to scale responsibly, blending generative AI with reliable, testable flows that can be audited and governed. With CALM, Grace can begin to support more personalized, context-aware interactions that reflect patient needs, coverage, and care status, opening the door to a more unified experience across digital services.
“We’ve built a scalable, compliant platform with Rasa that supports real patient needs and drives measurable outcomes. CALM gives us a framework to keep that momentum going, with the flexibility to grow responsibly as expectations around AI continue to evolve," said Wayne Foley, Director Software Engineering, Digital Innovation Group, Providence.
