Why it matters
- 250K+ teams validates it as the leading visual conversation design platform for AI agents.
- Cross-functional design — product managers, UX designers, and engineers collaborate on the same canvas.
- Covers the full pipeline: design, prototype, A/B test, deploy, and iterate in one platform.
- Support for both voice (Alexa, Google) and chat (web, Slack, Teams) in a unified canvas.
Key capabilities
- Visual canvas: Drag-and-drop conversation flow builder — no coding required for most workflows.
- AI responses: LLM-powered dynamic responses (GPT-4, Claude, others) within conversation flows.
- Knowledge base: RAG-powered knowledge base — upload docs, and the AI answers from them in context.
- API integration: Connect to any REST API with no-code request builders.
- Code blocks: JavaScript for custom logic when no-code isn't enough.
- Intent classification: Traditional NLU-style intent matching alongside LLM responses.
- Multi-channel: Deploy to web chat widget, Slack, Teams, Alexa, Google Assistant, phone (Twilio).
- Prototyping: Test conversations with simulated users and share previews.
- Collaboration: Real-time multi-editor canvas with comments, version history, and branching.
Technical notes
- Canvas: React-based visual flow editor
- AI: GPT-4, Claude, and custom model integrations
- NLU: Voiceflow NLU + LLM classification hybrid
- Channels: Web widget, Slack, Teams, Alexa, Google Assistant, Twilio (voice)
- Code: JavaScript code blocks for custom logic
- Pricing: Free (2 editors); Team ~$50/editor/mo; Enterprise custom
- Founded: 2019 by Braden Ream, Michael Hood, and Andrew Lawrence; Toronto; raised $90M
Ideal for
- Product and design teams building AI customer service or internal assistant experiences with visual tools.
- Organizations deploying across both voice (Alexa, phone) and chat (web, messaging) channels from one platform.
- Teams who want design, prototyping, and deployment in one tool without separate tooling for each phase.
Not ideal for
- Simple document chatbots where Chatbase or CustomGPT are faster and cheaper.
- Developers who prefer code-first approaches — Rasa or LangChain give more control.
- Very simple chatbots with basic FAQ responses — the platform's power is wasted on simple use cases.