Rasa is an agent platform for product teams building customer-facing AI. Teams start with a prompt and tune it in a visual interface, in code or with a coding agent, then add order and conditions where they want control. Rasa evaluates the agent on a hundred simulated conversations in voice and text, supports end-to-end checks and version comparison before release, and reads live conversations to tag and cluster issues. Deployment can run on your own infrastructure and model, and Rasa was named a Strong Performer in The Forrester Wave 2026 for conversational AI for customer service.
It is aimed at enterprise product teams, conversational AI groups, developers and IT directors in industries such as finance and banking, insurance, telecom, retail, healthcare, travel and the public sector. Non-developers can build, review and monitor agents in the Studio workspace, while engineers work in a version-controlled codebase from their IDE using the CLI and SDKs. Organizations with data-sovereignty requirements can keep their stack, data and models under their own control.
Typical uses include customer support and customer experience agents, sales enablement and operational efficiency projects. One agent can answer a hotline, Slack, web chat, WhatsApp and email, and teams can consolidate several bots into one assistant. A customer story in the page reports fallback rates falling from 18% to 3.5% after consolidating five bots, with expansion to voice. Other reported outcomes include an 80% contact reduction in retail and a 50% operational cost reduction in telecom. Rasa MCP Tools let coding agents work in the shared codebase to automate builds, tests and reviews.
