Intelligent workflow automation for business teams

Workflows are multi-step automations that combine triggers, AI reasoning, and actions across your tools. They start from an event—like a new email, form submission, calendar event, or scheduled time—and execute a series of steps you define. At each step, Workflows can pull context from your Knowledge Base, apply logic, generate content, update records, send messages, or route data. Unlike rigid scripts, Workflows can reason through complexity and handle exceptions based on the rules you set.

No. Cassidy's no-code builder lets you create Workflows by dragging and dropping actions, configuring inputs, and connecting triggers—all without writing code. The Workflow Copilot helps you structure logic and refine steps using plain language. You can start from templates or build from scratch.

Workflows can start from events in connected tools—new emails, form submissions, calendar events, Slack messages, CRM updates, helpdesk tickets, and more. You can also trigger them via webhook, run them on a recurring schedule, or start them manually.

Workflows support 100+ pre-built actions: generate content, search your Knowledge Base, send emails, update Salesforce or HubSpot records, create documents, post Slack messages, route data based on conditions, loop through lists, and more. You can also connect to any service via webhooks and API.

You can insert approval steps anywhere in your Workflow. These pause execution until a person reviews the output, provides input, or approves the next action. Use them for quality checks, sensitive decisions, or anywhere human judgment is required before proceeding.

Workflows can pull from your Knowledge Base at any step. Cassidy retrieves relevant documents, policies, or data to inform AI reasoning and generate accurate outputs. Every AI-generated response cites the source it used, so results are traceable and verifiable.

Cassidy supports models from OpenAI, Anthropic, and Google. You select the model for each step based on the task—writing, reasoning, analysis, or research. New models are added as they become available, and you can switch without rebuilding your Workflows.

Yes. You can set permissions at the Workflow level to control who can view, edit, run, or deploy. Share org-wide, limit to specific teams, or keep Workflows private. Knowledge Base permissions also apply—Workflows only retrieve information users are authorized to access.

No. Your data is never used to train AI models. It's used only to execute your Workflows and is encrypted in transit and at rest.