Introduction

Use your own model provider key to ask questions and perform permitted workspace actions in Blue.

The AI Agent is a chat assistant in Blue. It can answer workspace questions and perform permitted actions from a conversation.

The AI Agent is currently available on all plans. Your organization must provide an API key from Anthropic, OpenAI, Google, DeepSeek, or OpenRouter. Your model provider bills the key’s usage directly.

What the AI Agent Can Do

  • Answer questions about records, lists, activity, and Blue documentation.
  • Search and read your workspace Docs to answer questions about your organization’s own processes, policies, and notes. It searches the current workspace by default, and every workspace you can open when you ask it to.
  • Create and update records, set custom field values, and add tags, lists, checklists, and comments.
  • Apply filters, grouping, and saved views from the AI Agent panel.
  • Use three autonomy modes. See Using the Assistant.
  • Require approval for applicable bulk, destructive, non-reversible, or uncertain actions. See Safety and Approvals.
  • Follow organization prompts, workspace prompts, and reusable skills.

The AI Agent uses the requesting person’s existing Blue permissions. It cannot use the assistant to bypass workspace or record access.

Start Using the AI Agent

  1. Ask an organization owner or administrator to configure a provider key under Settings > AI > General. See Getting Started.
  2. Set access to all members or specific members. See Access Control.
  3. Select AI Agent at the bottom of Blue.
  4. Start a conversation. See Using the Assistant.