Relevance AI

Relevance AI is a no-code AI agent builder where sales, support and operations teams combine agents into workforces, on flat plans metered in Actions.

Relevance AI turns the repetitive parts of a team's work, such as researching leads, answering tickets or preparing meeting briefs, into AI agents that run on their own. OnSearch Pty Ltd, an independent Australian company trading under the Relevance AI name, builds and sells the platform. Its team works from Sydney, San Francisco and Barcelona. The aim is an AI workforce that takes on routine tasks so people can focus on the work that needs them. An agent starts from a prompt, a language model and a set of Tools. Each agent can run on its own model, with a fallback model from another provider for when the first one fails. Tools are step-by-step actions, Python code steps among them, and Knowledge holds the documents an agent retrieves from at run time. The Workforce Builder places several agents on a canvas so they delegate work to one another, with approval steps where a person signs off. Invent, the built-in copilot, takes a plain description of an agent and wires up its tools, knowledge and evaluations. The platform is aimed at go-to-market, support and operations teams whose subject-matter experts want to automate their own processes. Smaller groups start on the self-serve Pro plan, and the Team plan adds shared projects, end users and calling and meeting agents. Enterprises buy through sales and get single sign-on, audit logs, Agent Evaluations and an embedded deployment team that maps the first use cases. Every plan meters usage in Actions, one per tool run, and Vendor Credits, which cover model costs. Sales teams run agents that research inbound leads, write pre-meeting briefs and nurture qualified prospects toward a demo. Support teams start agents from new Zendesk tickets, embed one on a website as a chat widget, or let it answer in Slack and WhatsApp. Developers call a finished agent over HTTP through its API trigger, or reach it from Claude Desktop and Cursor through the Relevance AI MCP server. The Tasks page shows every run as a timeline, and version history rolls an agent back when a change breaks it. Operations teams build on the same blocks for back-office work such as invoice matching and contract review.