AI Planning

Planning is where most projects fail, not execution. And the problem isn't lack of ideas, it's lack of challenge. Teams jump straight into building without validating assumptions, without challenging scope, without identifying obvious risks.

AI Planning in Almirant isn't a tool that generates work items. It's a sparring partner that challenges you before you write a single line of code.

Why a sparring partner, not a generator

The difference is fundamental:

• A generator takes your input and produces output. You say "I want authentication" and it gives you 15 tasks.
• A sparring partner challenges your input. It asks why you need authentication, what alternatives you considered, where the risks are.

The result isn't just a list of items: it's validated scope and challenged assumptions.

From vague idea to validated scope

From vague idea to validated scope in one session. Auto-generated epics, stories, and tasks with definitions of done.

Planning Sessions

Planning sessions are multi-turn conversations with the AI. It's not a single prompt: it's an iterative dialogue where the AI:

• Challenges assumptions -- Questions decisions you take for granted.
• Identifies risks -- Points out blind spots in your planning.
• Proposes alternatives -- Suggests approaches you hadn't considered.
• Refines scope -- Helps define what's in and what's out.

Session states

Each session goes through different phases:

Phase	Description
idle	Waiting for you to start the conversation
chatting	Exchanging messages with the AI
streaming	Receiving response in real time
thinking	AI processes and reasons (thinking mode active)
reviewing	Reviewing suggested items to accept or reject
completed	Session finished
paused	Session paused to continue later

Session history

Your sessions are automatically saved and you can access them later:

• Active -- Session in progress, you can continue the dialogue.
• Completed -- Session closed with items created.
• Archived -- Session saved for future reference.

Each session records tokens used (totalInputTokens, totalOutputTokens), estimated cost, and duration, so you can optimize your quota usage.

Ideation seeds

Before starting the planning dialogue, you can select seeds: previous ideas captured in Almirant that serve as a starting point. Seeds provide initial context to the session:

• Saved brainstorming notes.
• Ideas from previous sessions.
• Requirements captured from stakeholder conversations.

Workflow

1. Select seeds (optional) -- Choose previous ideas to provide initial context.
2. Start the session -- Describe what you want to achieve, but expect to be challenged.
3. Dialogue with the AI -- Answer questions, defend decisions, reconsider assumptions.
4. Review the proposal -- When the scope is clear, the AI generates structured work items.
5. Accept or reject items -- Select those that apply, discard those that don't.
6. Items created on the board -- Accepted items are automatically created on the active board.

Writing effective prompts

The conversation is bidirectional, but your first message matters:

Good starting points

Prompt	Why it works
"I want to add social auth, but I'm not sure if it's worth the effort vs magic links"	Invites the AI to compare alternatives
"I need to improve product listing performance, users are complaining about slowness"	Defines the real problem, not the assumed solution
"We need to migrate from REST to GraphQL, the team says it's better but I'm not convinced"	Opens space to challenge the decision

Avoid

Prompt	Problem
"Generate tasks for authentication"	Closes dialogue space, asks for direct output
"Build a user CRUD"	No problem to solve, just mechanics
"Improve the whole system"	No focus, impossible to challenge

Response modes

The AI can operate in different modes:

• Streaming -- You see the response as it's being generated, ideal for rapid iteration.
• Thinking -- The AI reasons step by step before responding, better for complex problems.

You can alternate between modes depending on the complexity of what you're discussing.

Reviewing suggestions

When the AI generates its work item proposal, you'll see a list organized by type:

• Epic -- A large initiative that groups multiple features.
• Feature -- A concrete functionality that delivers value.
• Story -- A user story with acceptance criteria.
• Task -- A specific, actionable technical task.

Each item includes title, description, and proposed hierarchy. You can:

• Accept individual items.
• Reject items that don't apply.
• Accept all if the entire proposal looks good to you.

Prerequisites

To use AI Planning you need:

• An AI provider configured with a valid API key.
• Available quota in your organization.
• A project with at least one active board where the items will be created.

For Developers

MCP Tools

The following tools are available via MCP to interact with planning sessions:

Tool	Description	Main parameters
record_ai_session	Records an AI planning session	projectId, prompt, response, tokensUsed
get_ai_sessions	Retrieves the AI session history for a project	projectId, limit

Example: Recording a planning session

Tool: record_ai_session
Parameters:
  projectId: "project-uuid"
  prompt: "I need a push notification system"
  response: "AI-generated proposal..."
  tokensUsed: 1250