Don't outsource your intelligence

Build smarter bots that can handle multi-turn, contextual conversations. Keep your labeled training data to yourself and be independent of 3rd party APIs.

Case Study

Tia logo

HIPAA compliant natural language understanding for a digital health assistant, allowing users to get their top health questions resolved while their data is secure


Tia, a San Francisco-based healthcare startup, allows its users to answer health questions via natural language. The conversation happens within their own app, which has been downloaded thousands of times.


Being in the healthcare space, HIPAA compliance is very important. For this reason, Tia would have had to build their own natural language understanding AI. However, developing this from scratch takes months and requires specific AI and Machine Learning talent.


Tia decided to build their conversational natural language understanding on the Rasa Stack, which is actively used by thousands of developers worldwide. Extending this with the Rasa Platform for faster NLU training, Tia was able to set up in a few days what would have taken months. Adding more training data from real users makes the system more robust over time and increases the quality.

Next Step

Extension to a wider audience and the implementation of more skills to support even more requests.

A demo of the Rasa integration

Bots and Digital Assistants for your industry

Want to build smarter bots? Rasa is made for you.

Easier to maintain and train

Forget rule-based systems. Manage your conversations with machine learning, and watch it improve with each conversation.

Integrates into your workflow

Rasa seamlessly integrates with your existing application and backends.

Handle complex conversations

Conversations between humans are rarely one question and one answer. Rasa's advanced dialogue management is based on machine learning and allows for smarter conversations.

Maximum control

Deploy Rasa on-prem or on your private cloud. Keep your valuable training data to yourself and tweak your machine learning models for the best performance.

Want to learn more?