How to use the Rasa Platform | Rasa Documentation

Build your first agent in just a few minutes with Rasa Copilot.

Rasa makes it easy for teams to build, test, and improve AI assistants through an iterative development process. Below, you'll find the key stages of the conversational AI development lifecycle and how Rasa can help streamline your workflows. With each iteration, you can refine your assistant based on real user insights, ensuring continuous improvement and better business outcomes.

Workflow with the Rasa Platform

1. Build

Design, develop, and configure user journeys while improving existing ones. Use these platform tools to streamline development:

Write Content Studio

• Create, edit, and delete templated responses centrally in the Rasa Studio CMS.
• Create, edit, and delete intents directly within the CMS when using Rasa NLU.
• Empower copywriters, designers, and AI trainers to maintain conversational content efficiently.

Build Flows Studio

• Visually build conversational experiences using the flow builder in Rasa Studio.
• Structure the business logic of customer interactions into modular, reusable flows.
• Combine different elements to build everything from simple Q&As to complex, multi-turn conversations—without writing code.

Train Models Studio

• Train your assistant to apply updates after defining flows and content.

Integrate Channels Pro

• Configure and manage channel integrations in Rasa Pro.
• Add and maintain integrations across iterations, making modifications as needed.

Configure Models Pro

• Integrate and fine-tune language models to optimize dialogue management.
• Enhance responses using retrieval-augmented generation (RAG) services.

Build Actions Pro

• Write custom Python actions to handle complex logic, integrate APIs, and extend conversational capabilities.
• Surface custom actions in Rasa Studio for reuse across your assistant.

2. Test

Verify and refine your assistant’s behavior before deployment.

Inspect and Debug Studio

• Use Rasa Studio’s Inspector to test and debug the conversation flows you built.
• Visualize flow navigation, track collected information (slots), and analyze responses.
• Quickly identify issues and refine conversations for a smoother user experience.

Run Automated Tests Pro

• Execute automated tests written in Rasa Studio or Rasa Pro.
• Validate end-to-end assistant behavior to ensure consistency and reliability.
• Use automated testing to verify integrations and prevent regressions.

3. Deploy

Promote your tested assistant into production using your preferred infrastructure.

Deploy Pro

• Deploy in a scalable, production-ready environment with Rasa Pro.
• Choose on-prem, cloud, or Kubernetes deployment options.
• Leverage built-in versioning, rollback support, and monitoring for smooth rollouts.

4. Review

Analyze real conversations and refine your assistant based on user interactions.

Conversation Review Studio

• Access and analyze production conversation logs in the Conversation Review panel.
• Filter by topic or key metrics to identify patterns and optimize responses.
• Debug and refine assistant behavior using real-world data.

Monitor Performance Pro

• Use Rasa Pro’s monitoring tools for real-time visibility into your assistant’s performance.
• Track logs, analytics, and dashboards to optimize dialogue flows and detect issues early.

Next Steps:

• Set up your conversational AI team
• Start building a Rasa assistant for free