AI Assistant

Spedy includes an AI assistant that helps you manage your projects through natural conversation. Ask it questions about your boards, get morning briefings, create tickets, or let it suggest what to work on next -- all without leaving the chat interface.

Opening the AI assistant

Click the Chat item in the sidebar to open the AI assistant. You can start a new conversation or continue an existing one. The assistant remembers context within a conversation, so you can ask follow-up questions naturally.

What the assistant can do

Project briefings

Ask the assistant for a summary of what has changed since your last visit. It will look across all your boards and tell you about new comments, status changes, assignments, and any blockers that need attention.

Example prompts:

• "What happened since I was last online?"
• "Give me a briefing for today"
• "Are there any blockers I should know about?"

Create and edit tickets

Describe what you need in plain language and the assistant will create a ticket for you -- complete with a title, description, acceptance criteria, and estimated hours. It works in any language.

Example prompts:

• "Create a ticket for fixing the login timeout issue on the CS board"
• "Add a feature request for dark mode support"

Get suggestions

The assistant can recommend what you should work on next based on priority, due dates, and your current assignments. It considers open blockers and team workload to give you actionable advice.

Example prompts:

• "What should I work on next?"
• "Which tickets are overdue?"
• "Show me my open assignments"

Manage tickets through chat

You can ask the assistant to move tickets between statuses, change assignments, set priorities, or update due dates -- all through conversation. The assistant will confirm actions before executing them, so you stay in control.

Example prompts:

• "Move CS-42 to In Progress"
• "Assign the login bug to Sarah"
• "Set the priority of APP-17 to Critical"

Deep Research mode

For complex questions that require analyzing multiple boards and tickets, toggle on Deep Research mode. The assistant will perform a more thorough analysis with multiple rounds of context gathering before responding.

Specification generation

Deep Research can also generate detailed ticket specifications. When you ask the assistant to create a specification for a ticket, it uses Deep Research to gather context from your boards, related tickets, and project documentation, then produces a structured specification document. The generated specification can be applied directly to the ticket description.

AI Knowledge Base

On the Pro plan, you can build a Knowledge Store that the AI assistant draws from when answering questions. Upload documents, guidelines, architecture overviews, or any reference material and the AI will use them as additional context during conversations.

The Knowledge Base is managed from Settings > AI Knowledge Base. Documents are processed, chunked, and stored as vector embeddings so the assistant can perform semantic search across your knowledge store when formulating responses.

Task Assist

When creating a ticket, you can use Task Assist to generate a structured ticket from a short description. Write a brief sentence about what needs to be done and the AI will produce:

• A clear title
• A detailed description
• Acceptance criteria
• Estimated hours

Task Assist works in any language -- describe the task in German, English, or any other language and it will generate the ticket content in the same language.

AI data processing consent

Before the AI chat can be used, an administrator must grant AI data processing consent once for the entire organization. Chat input and context data are forwarded to Grok (xAI).

• The first time anyone opens the AI chat, a dialog explains what data is transmitted and why.
• Only admins can grant consent. Non-admins see a notice asking them to contact an admin.
• Consent can be revoked by an admin at any time -- AI chat features are then immediately disabled for all members.
• Only the data required for the specific feature is transmitted (data minimisation).

How it works behind the scenes

The AI assistant uses a multi-step process:

1. Thinking: Your question is analyzed and optimized for better results
2. Context gathering: The assistant searches your boards, tickets, and team data for relevant information
3. Response: A final answer is generated using the gathered context, with interactive actions when applicable

The assistant only has access to data you can see -- it respects your board permissions and role assignments.

Usage limits

AI usage is metered by tokens consumed per month, not by request count:

Token usage resets monthly. You can monitor your current consumption in your organization settings. Upgrading to Pro unlocks the AI assistant, Deep Research, the AI Knowledge Base, and all AI-powered features across the platform.

Note: MCP integration is free on all plans and does not consume AI assistant tokens. MCP allows external AI tools to connect to Spedy, which is separate from the built-in AI assistant.