Dialogflow is a 📝Google product for building conversational interfaces — text chatbots and voice agents — on top of 📝natural language processing that maps what people say to intents and actions. It lets you connect with users on your website, mobile app, the Google Assistant, Amazon 📝Alexa, Facebook 📝Messenger, and other popular platforms and devices. It began as API.AI, launched in September 2014 by the Speaktoit team behind a popular voice assistant app; Google acquired it in September 2016, renamed it Dialogflow in October 2017, and folded it into Google Cloud, where it split into the original ES edition and the flow-based CX edition. CX was rebranded as Conversational Agents in 2024–2025, and Google now positions CX Agent Studio, part of Gemini Enterprise for Customer Experience, as its successor while ES and CX remain supported; its closest rival is 📝Amazon Lex. It serves developers and enterprise contact centers, billed per request through Google Cloud with a no-cost trial edition, and third-party front ends such as 📝Botcopy wrap it in branded web chat.
Dialogflow was the language layer under 📝BrianBot's chatbot era: 📝ManyChat ran the broadcasts while 📝Janis caught the replies the bot couldn't parse and trained Dialogflow in the process. In June 2019 I used the same stack to spin up a bot for 📝Spiritual Bro customer (bro) service.
Key Features
- Intents and entities — Matches what a user says to a defined intent and extracts parameters like dates, places, and numbers using built-in system entities or custom ones.
- Visual flow builder — Dialogflow CX models conversations as pages and flows in a state-machine console, keeping large agents with many branching paths manageable for teams.
- Multichannel integrations — Connectors and an API deploy one agent to web chat, phone and IVR, messaging apps, and voice assistants without rebuilding its conversation logic.
- Fulfillment webhooks — Calls your own backend service when an intent matches, so the agent can look up orders, book appointments, or write records before composing a reply.
- Generative playbooks — Newer agents pair deterministic flows with LLM-driven playbooks and data stores, answering open-ended questions grounded in your own documents and sites.
Getting Started
- Sign in to the Google Cloud console and create or select a project with billing enabled.
- Enable the Dialogflow API, then open the Dialogflow CX or ES console for that project.
- Create an agent, choose its default language and region, and add your first intents.
- Test the agent in the built-in simulator, then connect a channel integration to deploy.
Related
- 📝Chatbot Development — the broader practice of designing and building conversational agents