Skip to main content

​
What is project chat?

Each project in Toots includes a collapsible chat panel where you can have an ongoing conversation with the AI about your project. The AI can:
  • Ask clarifying questions to understand your needs
  • Generate new tickets based on additional context
  • Update existing tickets (priority, description, status, etc.)
  • Remove tickets you no longer need
  • Provide recommendations on project scope and approach

​
Opening the chat

The chat panel is available on every project page. Click the chat toggle or panel to expand it alongside your Kanban board.

​
How the AI chat works

The AI assistant uses specialized tools to interact with your project and tickets. Understanding these tools helps you make better requests.

​
Available AI tools

The AI has access to five tools:

​
setClarifyingQuestions

When the AI needs more information, it uses this tool to present 2-4 clarifying questions. Questions may include multiple-choice options to make answering easier. Example questions:
  • “Who is the target audience for this feature?”
  • “What’s the expected timeline?” (with options like “1 week”, “2-4 weeks”, “1+ months”)
  • “Are there any technical constraints we should consider?“

​
listTickets

The AI uses this tool to find ticket IDs when you refer to tickets by name. For example, if you say “remove the onboarding ticket,” the AI first calls listTickets to find the ticket ID, then performs the requested action. Returns: Ticket ID, title, status, type, and priority for each ticket in the project.

​
updateTickets

This tool allows the AI to modify one or more tickets. It can update:
  • Title
  • Type (Story, Task, Epic, Milestone, Deliverable)
  • Priority (P0, P1, P2, P3)
  • Description
  • Status (todo, in-progress, done)
  • Estimated effort (XS, S, M, L, XL)
  • Acceptance criteria
  • Dependencies
  • Labels
Example request: “Make the user authentication ticket P0 priority”

​
removeTickets

The AI can delete one or more tickets by ID using this tool. Example request: “Remove the social login ticket—we’re not doing that anymore”

​
generateTickets

When the AI has enough context (from your initial description and clarifying questions), it uses this tool to generate a complete set of tickets for your project. Each ticket includes:
  • A unique ID (e.g., “ticket-1”, “ticket-2”)
  • Title and description
  • Type, priority, and estimated effort
  • Acceptance criteria (2-3 items minimum)
  • Dependencies on other tickets
  • Relevant labels
Tickets are created in logical order and stored in the database with an initial status of “todo”.

​
Common chat workflows

​
Refining project scope

1

Describe what you want to change

Tell the AI how you want to adjust the project scope.Example: “Actually, we don’t need the admin dashboard—let’s focus on the core user experience.”
2

AI updates tickets

The AI uses listTickets to find relevant tickets, then removeTickets to delete tickets that are no longer needed, or updateTickets to adjust priorities and descriptions.
3

Review changes

Check your Kanban board to see the updated ticket list.

​
Modifying ticket details

1

Request the change

Ask the AI to update specific ticket properties.Examples:
  • “Make the API integration ticket P1 instead of P2”
  • “Add a new acceptance criterion to the onboarding ticket: ‘User can reset password’”
  • “Change the effort estimate on the database migration to XL”
2

AI finds and updates the ticket

The AI calls listTickets to identify the ticket, then updateTickets with the specific changes.
3

Verify the update

Open the ticket detail sheet to confirm the changes were applied correctly.

​
Adding more tickets

1

Describe new requirements

Tell the AI about additional work that needs to be done.Example: “We also need to add email notifications when a user completes their profile.”
2

AI asks clarifying questions (if needed)

The AI may use setClarifyingQuestions to understand the requirements better before generating tickets.
3

AI generates new tickets

The AI calls generateTickets with the updated project context to create additional tickets.

​
Removing tickets

1

Specify what to remove

Tell the AI which tickets to delete.Example: “Remove the social media integration tickets—we’re cutting that from scope.”
2

AI identifies and removes tickets

The AI uses listTickets to find matching tickets by title or description, then calls removeTickets to delete them.
3

Confirm removal

Check your board to ensure the correct tickets were removed.

​
Tips for effective AI conversations

Be specific: Instead of “update the login ticket,” say “change the login ticket priority to P0 and add acceptance criteria about password validation.”
Reference tickets by name: You don’t need to know ticket IDs—the AI will look them up. Just use the ticket title in your request.
Provide context: If you’re changing direction, explain why. This helps the AI make better recommendations and generate more appropriate tickets.
Ask for clarification: If you’re unsure about something, ask the AI for recommendations or explanations.

​
Chat history

Your conversation history is preserved for each project, so you can refer back to earlier discussions and the AI maintains context across multiple sessions.

​
Next steps

Managing tickets

Learn about manual ticket editing and management

Exporting tickets

Send your tickets to Jira or Linear (coming soon)

Build docs developers (and LLMs) love

Get started for freeTalk to us