About Kyruus
Kyruus Health is a leading provider of care access solutions, powering more than 145 million patient searches annually. Their flagship platform, Kyruus Connect, helps over 1,400 hospitals, 550+ medical groups, and 100+ health plans connect more than 500,000 providers to patients seeking the right care.
Their mission: simplify healthcare navigation and improve access through better search, scheduling, and provider data management.
Kyruus’s Challenge
Traditional healthcare search experiences are riddled with friction. Patients using checkbox-based interfaces must translate everyday symptoms into medical jargon—filtering through provider lists that often return irrelevant or incomplete results. Kyruus recognized the growing demand for a more intuitive, conversational way to search for care—one that would reduce abandonment, improve satisfaction, and meet the expectations set by modern GenAI tools like ChatGPT and Perplexity.
Why Tribe AI?
Kyruus partnered with Tribe to accelerate their generative AI transformation. After an initial connection through the Francisco Partners Conference, Kyruus’s CTO, Harshit Shah, saw Tribe as a strategic partner with deep technical expertise in foundation models and a track record of enterprise-grade AI delivery. Tribe offered not just architecture guidance, but the ability to pressure-test assumptions and reduce time to value without compromising quality.
“We were looking for a partner who brings what good looks like in terms of architecture patterns and could help us accelerate from POC to general availability.” – Harshit Shah, CTO, Kyruus Health
The Solution: Guide, a GenAI-Powered Patient Search
Together, Kyruus and Tribe built a conversational AI experience—Guide—that allows members to search for providers using natural language (e.g., “my knee hurts after climbing stairs”). This AI-powered system interprets intent, matches members with the right providers, and surfaces results in seconds.
Unlike legacy filters, Guide provides results that are:
- Context-aware: Interprets the full nuance of a patient’s input.
- Location- and availability-specific: Aligns matches with real-world constraints.
- Conversationally surfaced: Reduces the need for medical translation.
How It Works
- Input: A patient enters a query via Kyruus’s web or mobile app.
- Processing: The query is routed via API Gateway to a custom conversation service powered by Claude 3.5 Sonnet on Amazon Bedrock.
- Understanding: The LLM extracts clinical meaning and converts it into structured parameters—such as specialties, procedures, and insurance filters.
- Retrieval: These parameters are used to query a provider database via Amazon OpenSearch.
- Output: Claude generates a conversational recommendation that explains the chosen parameters clearly and accurately, while also suggesting follow-up questions.
This end-to-end experience reduces decision fatigue and eliminates the need for medical translation on the patient’s part—returning high-quality, relevant results in just one page.
Tech Stack Details
- Foundation Model: Claude 3.5 Sonnet (via Amazon Bedrock)
- Search Engine: Amazon OpenSearch
- Orchestration: Amazon ECS & API Gateway
- Storage: Amazon S3
Kyruus’s Experience Working with Tribe
“We didn’t want to build another chatbot. Tribe helped us deliver a truly integrated experience.” – Harshit Shah, CTO, Kyruus Health
Kyruus describes the engagement as one of the most collaborative and impactful partnerships in their AI journey. From initial discovery sessions that challenged incorrect assumptions, to daily standups and rapid demos, Tribe embedded tightly with the Kyruus engineering team. Their SMEs provided critical guidance on everything from testing infrastructure to hallucination mitigation and model validation.
According to Harshit, the project could have taken 2-3x longer without Tribe’s involvement—and may have resulted in brittle or outdated architecture without their foresight. “Partnering with Tribe allowed us to shorten time to market while building in-house expertise,” said Shah. “Instead of taking 4-6 months, we got there in two.”
Tribe Team Members
- Technical Lead: Kashyap Murali
- AI Engineer: Shalini Ananda
- AI Engineer: Nishith Agarwal
The Impact & The Future
With Guide now in production, Kyruus is seeing measurable improvements across every stage of the patient journey. “Our conversion rate for members scheduling appointments went up by more than 20%—they’re finding what they need faster and more reliably,” said Shah.
Early outcomes show significant gains across the board:
- 1.5 seconds to return provider results
- 75% increase in provider match satisfaction
- 40% faster appointment scheduling
- 60% reduction in search abandonment
- 35% decrease in call center volume
- 42% drop in no-shows
- 31% increase in accurate specialty referrals
Kyruus is now positioned to continue evolving Guide—with the flexibility to test new models, expand functionality, and further personalize the member experience. What started as an AI-powered search upgrade is becoming the foundation for a smarter, more connected care navigation system.
