What challenges did Kakao Healthcare face before AI implementation?

Kakao Healthcare needed to deliver faster, smarter, and more personalized healthcare services through an AI-driven app to support chronic disease management and individual health journeys. Managing complex technical infrastructure while ensuring personalized, accessible, and secure healthcare services was a core challenge.

What was the impact of the AI solution?

Deployed a more accurate agentic AI chatbot in a few weeks, enhancing personalized care and user experience. Improved healthcare-related answers with faster response times and better accuracy. Implemented robust security features to guard against misuse and ensure safety. Set up a scalable AI architecture with plans for expanding specialized agents and voice interaction.

Source

GCPProductionEvidence: Medium65/100

Kakao Healthcare deploys agentic AI chatbot for personalized healthcare app using Google Cloud AI

Use case typePatient engagementUpdated Jun 13, 2026

Kakao Healthcare developed PASTA, an AI-powered healthcare app, to provide personalized care and smarter health management. The app uses an agentic AI chatbot powered by Google Cloud's Vertex AI and Gemini models, integrated with multi-layered agents for health, app support, and security. Features include Vision AI for food recognition, speech-to-text health logs, and lifestyle change notifications based on personal health data. The chatbot utilizes classification modules, rewrite modules, Vertex AI, Gemini Flash, and a context cache in Firestore for persistent and relevant user conversations. Security layers protect against prompt breaking and hijacking, ensuring safe usage. Future plans include expanding specialized AI agents and adding voice interaction capabilities for enhanced user experience.

Organization: Kakao Healthcare
Industry: Healthcare
Location: South Korea

Reported outcomes

Strategic outcomes

New product / capabilityDeployed an agentic AI healthcare chatbotCustomer experience & trustImproved personalized care and user experienceRisk & complianceAdded security guardrails against misuseScale & capacityBuilt a scalable AI architecture

Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks

Customer: Kakao Healthcare
Provider: GCP
Maturity: Production
Linked source: Google Cloud Customer Stories

Deployed a more accurate agentic AI chatbot in a few weeks, enhancing personalized care and user experience

Customer identity supportedSource describes one deploymentMaturity supported

Primary read

Use case focus

Showing 3 of 3

1AI Agents
2Conversational AI
3Healthcare Personalization

Kakao Healthcare needed to deliver faster, smarter, and more personalized healthcare services through an AI-driven app to support chronic disease management and individual health journeys.
Managing complex technical infrastructure while ensuring personalized, accessible, and secure healthcare services was a core challenge.

They built PASTA, an AI-powered chronic disease management platform, with an agentic AI chatbot using Google Cloud technology.
The chatbot distinguishes user intents and translates natural language into clear requests, retrieving data via Vertex AI and Gemini Flash.
Multi-layered AI agents run simultaneously for health, app, website, and security queries to provide expert responses.
User chat history is stored in Firestore for context retention, enabling faster and more accurate answers.
Security guardrails prevent misuse of the AI chatbot.

Technologies

Vertex AI Gemini Firestore Model Garden Generative AI on Google Cloud

Deployed a more accurate agentic AI chatbot in a few weeks, enhancing personalized care and user experience.
Improved healthcare-related answers with faster response times and better accuracy.
Implemented robust security features to guard against misuse and ensure safety.
Set up a scalable AI architecture with plans for expanding specialized agents and voice interaction.

Architecture

The AI chatbot uses classification, rewrite, and reranker modules along with Vertex AI and Gemini Flash models to process user queries. Firestore stores chat history for context persistence. Multi-layered agents handle health, app, website, and security queries. Security features protect against prompt breaking and hijacking.

Sources & evidence2

Evidence: Medium65/100Evidence strength