This article covers how to interact with Agents once they're set up, along with example Agents you can use as a starting point.
👉 Not sure what Agents are or how to set one up? Start with AI Agents: Overview, and Setting Up and Managing AI Agents.
What Can Agents Do?
Agents can access most areas of Productive — from tasks and projects to reporting, finance, and people management. What a specific Agent can do is determined by the permission set assigned to it.
Permission sets work the same way for Agents as they do for users — the more access an Agent has, the more it can do. An Agent with limited permissions might only be able to read and comment on tasks, while one with broader access could update records, adjust budgets, or manage time entries across projects.
📌 Agents are private by default — only their creator can see and use them. To make an Agent available to others, you'll need to share it. Learn how in Setting Up and Managing AI Agents.
Working With an Agent
Once an Agent is set up, there are three ways to work with it:
1. Talk to the Agent via the Assistant
Open the AI Assistant and select the Agent you want to work with. The Agent brings its instructions and knowledge into the conversation, but you're still the one taking actions.
Example:
Ask it to create a task - Everything created by the Agent is created under your name, but the Agent guides how it's structured (format, subtasks, description, etc.)
Ask it a question - It responds based on its instructions and the context it has access to
Ask it to review something - It checks the relevant tasks, docs, or projects and reports back
2. Triggers (Scheduled runs)
The Agent runs automatically at the time and frequency you define. You can add specific instructions for each trigger - what to check, what to do, and what to act on.
Example: Every Monday at 10:00, the Agent checks a project for milestones without an update in the past week and posts a follow-up comment to the relevant person.
3. Mention or Assign the Agent on a Task
Once you add an Agent as a project member, you can interact with it directly on tasks in two ways:
@mention it in a comment - The Agent will respond and act on whatever you ask, such as closing tasks, updating fields, or gathering information from docs.
Assign it to a task - The Agent monitors the task and decides whether to act based on what's happening, according to its instructions. You don't need to @mention it every time.
4. Sharing Agents
By default, Agents are private — only you can see and use them. To make an Agent available to others, open its profile in Resourcing > Agents, and click Share on the Agent's profile.
When sharing, you can choose an access type for each user:
Can view — the user can see and use the Agent
Full access — the user can see, use, and edit the Agent's settings
Example Agents
Documentation Helper - Helps you write and structure documents and specs. When assigned to a project, it can review task descriptions, suggest missing details, check for grammar issues, and help keep your docs consistent.
Researcher - Gathers information from your docs, tasks, and projects on a topic you specify and compiles a summary on demand. Useful for prep work, status overviews, or pulling together context before a meeting.
Weekly Standup Agent - Runs every Monday morning, checks a project for tasks without a recent update, and pings the assignees to report progress.
Notification Cleaner - Reviews your notifications, dismisses ones where you're not directly involved or mentioned, and lists those that need your attention.
Budget Watcher - Monitors project budgets and posts a comment to the budget feed, tagging the budget owner and project manager when spending crosses a threshold you define.
Project Assistant - When a new project is created, it checks that all required fields are filled in, creates a standard set of tasks, and assigns them to the right people based on their role.
