VALUTCARE

• 

  The Chat feature

• 

  Dashboard with Retool

• 

  Voice and TinyFish for searching on web

Inspiration

Healthcare clinics still rely on humans to answer phones and schedule appointments because AI systems cannot be trusted with sensitive patient data, making automation a compliance risk instead of a solution. We were inspired to build VALUTCARE to prove that autonomous, agent-driven systems can operate safely in regulated environments if privacy and governance are designed into the architecture from day one.

What it does

VALUTCARE is a voice-first, autonomous appointment scheduling system that handles an entire booking flow—intake, clinic discovery, time matching, confirmation, calendar creation, and notifications—while cryptographically ensuring that private health information never reaches the AI. Every decision is observable, auditable, and replayable through a governance control plane.

How we built it

We used Yutori to generate and execute a multi-step agent plan that adapts dynamically during the conversation, TinyFish to extract real-world availability signals from clinic websites, ElevenLabs to provide a natural voice interface, and Retool as a governance and compliance replay UI. PHI is detected, tokenized, and vaulted before any model call, and all sensitive execution happens strictly outside the orchestration layer.

Challenges we ran into

The biggest challenge was balancing autonomy with safety—ensuring the agent could act independently without ever being able to access or leak sensitive data. Designing clean trust boundaries, handling failures without breaking the plan, and making compliance visible (not just claimed) required careful system design under tight time constraints.

Accomplishments that we're proud of

We built a fully autonomous, end-to-end scheduling agent that completes real workflows while providing proof that PHI never entered the AI. We’re especially proud of the compliance replay experience, which allows anyone to inspect exactly what the agent saw, decided, and executed—turning a black box into a transparent system.

What we learned

Agentic systems become far more powerful and trustworthy when orchestration, execution, and data custody are deliberately separated. We also learned that observability and governance aren’t “enterprise extras”—they’re essential for deploying AI in real-world, regulated domains.

What's next for VALUTCARE

Next, we plan to integrate real EHR and calendar providers, expand beyond scheduling into intake and follow-ups, and formalize VALUTCARE as a compliance-first agent platform for healthcare operations. Long-term, we see VALUTCARE enabling a new class of autonomous systems that are safe enough to run where humans currently must.

Built With

• elevenlabs
• retool
• tinyfish
• yutori