Introduction

The AI Agent is a chat assistant built directly into Blue. It can answer questions about your workspaces, draft and edit records, manage checklists and comments, and carry out multi-step changes on your behalf — all from a conversation.

Unlike Blue’s other AI-assisted features, the AI Agent is bring your own key (BYOK): your organization connects its own API key from Anthropic, OpenAI, Google, or OpenRouter, and Blue uses that key to run the assistant. There’s no separate AI subscription — you pay your model provider directly for what you use, and Blue gives you usage tracking and a monthly spend cap to keep it predictable.

What it can do

• Answer questions about the records, lists, and activity in a workspace.
• Take actions — create or update records, manage checklists, add comments, and more — by calling the same tools Blue’s automations use internally.
• Work at three levels of autonomy, from read-only investigation to fully hands-off execution. See Using the Assistant.
• Ask for approval before higher-risk changes, including bulk updates and destructive actions. See Safety and Approvals.
• Follow custom instructions you define at the organization or workspace level, and reusable skills for repeatable tasks.

Getting started

1. An organization owner or admin enables the assistant and adds a provider API key under Settings > AI. See Getting Started.
2. Decide who in your organization can use it — everyone, or a chosen list of members. See Access Control.
3. Open the AI Agent panel from the sidebar and start a conversation. See Using the Assistant.