AI-powered changelog summarizer
Explore using an LLM to auto-generate weekly changelog summaries from merged PRs.
AI Agents
AI Feature Pipeline Board
Track AI features from idea to launch on one Kanban board
The AI Feature Pipeline Board gives product and engineering teams a structured way to move AI-powered features from early concept through prototyping, evaluation, and launch. Built on a five-column kanban board, it keeps every AI feature visible as it advances through technical spikes and testing. Manage it manually or let AI agents update tasks automatically through the MCP server and REST API, so your team always has an accurate view of what's being explored, tested, and shipped.
kanboard.io/templates/ai-feature-pipeline
Template preview
AI Feature Pipeline Board
5 columns
Concept
3Capture early-stage AI feature ideas before technical validation begins.
Smart task priority suggestions
Investigate ML model to suggest task priority based on due dates and activity.
Natural language task search
Concept for letting users search tasks with plain-language queries via the command palette.
Prototyping
3Build technical spikes and proof-of-concept implementations for approved concepts.
Prototype RAG-based support assistant
Build a retrieval-augmented generation prototype to answer user questions from docs.
MCP server auto-task creation spike
Test MCP server ability to auto-create tasks from external webhook events.
Keyboard shortcut AI command spike
Prototype AI-triggered actions accessible via keyboard shortcuts.
Evaluation
3Test prototypes with real data, gather metrics, and validate feasibility before launch.
Evaluate RAG assistant accuracy
Run accuracy and latency tests on the RAG-based support assistant prototype.
User test smart priority suggestions
Collect feedback from beta users on ML-based task priority suggestions.
Benchmark natural language search relevance
Measure relevance scoring of natural language task search results against manual search.
Launch Ready
2Track features that passed evaluation and are queued for release with docs and QA complete.
Finalize RAG assistant docs
Complete REST API documentation and macOS app integration notes for the RAG assistant.
QA sign-off for changelog summarizer
Final QA pass on AI changelog summarizer before shipping to production.
Shipped
2Archive AI features that have been released to all users.
Command palette AI search live
Natural language task search is now live in the command palette for all users.
MCP auto-task creation released
MCP server webhook-based auto-task creation is now available via the REST API.
How to run this board
Step 1
Add new AI feature ideas to Concept as tasks, describing the problem and proposed AI approach.
Step 2
Move validated concepts to Prototyping and build a technical spike using the MCP server or REST API.
Step 3
Shift prototypes to Evaluation to test accuracy, performance, and user feedback.
Step 4
Once evaluation passes, move tasks to Launch Ready for final QA and documentation.
Step 5
Move completed features to Shipped once released to users.
AI agent usage
Run it with AI agents
This playbook helps AI agents manage an AI Feature Pipeline board using the Kanban MCP server. Agents can create, move, and update tasks representing AI features as they progress from concept through evaluation to launch, keeping the board synced via MCP tools like create_task and move_task instead of manual edits.
## AI Feature Pipeline Board
This project uses a Kanban board with columns: Concept, Prototyping, Evaluation, Launch Ready, Shipped.
Agents should use the MCP server (list_tasks, create_task, move_task, update_task) to manage tasks.
When a feature idea is proposed, create a task in Concept. Move tasks to Prototyping once a technical spike starts, to Evaluation when testing begins, to Launch Ready when approved, and to Shipped once released. Always check list_tasks before creating duplicates. Respect the 25-task free plan limit.Move the 'RAG-based support assistant' task from Evaluation to Launch Ready and add a note about eval results.
List all tasks currently in the Prototyping column so I can prioritize this week's sprints.
Create a new task in Concept for an AI-powered changelog summarizer feature.
Learn more about the board-centric workflow in Kanban for AI Agents or open the MCP guide.
Frequently asked questions
How many AI feature tasks can I track on the free plan?
The free plan supports 1 project with up to 25 tasks total, which comfortably fits this 5-column AI Feature Pipeline board with room to spare.
Can AI agents manage this board automatically?
Yes, agents can use the MCP server tools such as create_task, move_task, and update_task to manage the pipeline without manual intervention.
Do I need the pro plan to use this template?
No, the free plan works fine for a single AI feature pipeline project. Upgrade to pro ($12/mo or $96/yr) if you need multiple projects or more than 25 tasks.
Can I customize the column names?
Yes, columns like Concept, Prototyping, Evaluation, Launch Ready, and Shipped can be renamed to fit your team's terminology while keeping the same workflow logic.
How do I move tasks between columns using the API?
Use the move_task MCP tool or the REST API to update a task's column and position as it progresses through the pipeline.
Is this board accessible from the macOS app?
Yes, the kanban board and all its columns and tasks are accessible from the macOS app as well as via the REST API and MCP server.
Related templates
Ready to create this board?
Sign in and use the template to create a project with columns and sample tasks.