Productive's AI Assistant lets you interact with Productive using natural language.
You can ask questions, generate outputs, and perform actions directly inside Productive without switching between tabs or setting things up manually.
📌 Important: AI-generated responses may be inaccurate or incomplete. Always review results before applying changes or relying on outputs.
Productive AI is available on all subscription plans and is free to use.
Important: AI features are subject to usage limits per user to ensure stable performance and fair use, which may evolve over time; if a limit is reached, some AI features may be temporarily unavailable—contact support if you need additional access.
Turning Productive AI On or Off
Admins control whether Productive AI is available for their organization. When Productive AI features are not enabled:
Users in the organization cannot access AI features
The organization’s data is not used to power AI functionality
To activate Productive AI:
Go to Settings > Productive AI
Toggle Turn on AI
Once enabled, all available AI features are added to the organization.
How the AI Assistant Works
The AI Assistant understands your requests and performs actions directly in Productive. When you ask it to do something, it might create a project, generate a report link, or draft an automation rule, depending on what you need.
Below are examples of what the Assistant can do at the moment.
Build Reports
The Assistant can generate reports based on natural language input. Instead of manually configuring filters and views, you can describe what you need in plain language.
Example: Check who's out this week
User input:
Show all users who are off this weekAssistant output:
A link to a bookings report with the correct filters already applied.
The Assistant generates interactive reports you can review, adjust, and share. You can also chat about the results or ask for a different view if needed.
Tip: In Reports, click the AI Report button in the upper right to enter your prompt or question. After submitting, the Assistant will appear on the left with a link to the report or follow-up questions to help you refine it.
📌 Permissions: Users can only generate reports based on data they have permission to access.
Manage Projects
The Assistant can handle core project and task operations. It can read existing data or create, update, and remove items within projects in Productive.
Supported actions include:
Creating and updating projects
Managing tasks
Working with docs
Example: Add a new project
User input:
Create a client project called Acme Rebranding Campaign for the Acme Client Company and set me as the project managerAssistant output:
A project is created with the specified name, company, project type, and project manager.
📌 Permissions: The Assistant respects your existing permission sets and cannot perform actions you are not allowed to do manually.
Build Automations
The Assistant can create automations based on natural language instructions. You can describe what should happen, and the Assistant will generate a draft rule with the correct triggers, conditions, and actions.
Example: Task triage
User input:
Assign every newly created task in the “Review” status to meAssistant output:
A draft automation with:Trigger: Task created
Condition: Task status = Review
Action: Update task → Assigned to = Me
Automations created by the AI are created as drafts and can be reviewed or edited before activation.
📌 Permissions: Only users with permission to create or edit automations can use this agent.
Tip: Clicking the AI Automation button in Settings > Automations opens the AI Assistant interface, where you can describe what you want to achieve and have the Assistant help build the rule with you.
Using the Start AI Agent Action in Automations
Automations can include a Start AI Agent action, which uses the AI Assistant as part of a workflow.
When the automation runs, it starts a background session with the Assistant and sends the prompt you defined. The Assistant understands the automation context, including what triggered it and which task, project, or item is involved.
You describe what should happen using natural language, instead of configuring every step manually.
Example prompt:
Add a comment to the task mentioning the project manager and asking them to share their thoughts.
Actions are executed in the automation creator’s name and follow their permission scope.
Automations that include this action are private by default (marked with a padlock icon 🔒) and visible only to the creator. This protects your AI-generated automation logic from being modified by others.
📌 Permissions: AI actions run within the automation’s permission scope and follow the user’s access rules.