Skip to content

AI Workflows

Learn how to effectively use AI assistants to manage your projects.

The AI-First Approach

graph LR
    A[Think] -->|Natural Language| B[AI Assistant]
    B -->|MCP Tools| C[Obsidian Vault]
    C -->|Plugin| D[Visual Review]
    D -->|Insights| A

    style A fill:#e1f5ff
    style B fill:#fff3e0
    style C fill:#f3e5f5
    style D fill:#e8f5e9

With this system, you think and communicate in natural language, and the AI handles the structured data management.

Common Workflows

1. Project Initialization

Goal: Start a new project from scratch

I'm starting a new project to build [description]. Help me structure it into milestones and stories.

Requirements:
- [requirement 1]
- [requirement 2]
- Target: [timeline]

AI will: 1. Ask clarifying questions 2. Suggest milestone breakdown 3. Propose initial stories 4. Create the structure

You then: 1. Review on canvas 2. Refine with AI 3. Add details

2. Daily Standup

Goal: Quick status update and planning

What did I work on yesterday? (check completed tasks)
What's in progress? (check InProgress items)
What should I work on today? (check priorities and blockers)

3. Sprint Planning

Goal: Plan next iteration

Show me all NotStarted stories in M-001
Which stories are highest priority?
What are the dependencies for S-005?
Create tasks for S-005 based on the description

4. Dependency Analysis

Goal: Understand what's blocking progress

What's blocking M-001 from completion?
Show me the dependency chain for S-003
Are there any circular dependencies?
What can I work on that has no blockers?

5. Progress Reporting

Goal: Generate status reports

What's the completion percentage of M-001?
List all completed stories this week
Show me stories assigned to [person]
What's at risk of missing the deadline?

6. Bulk Operations

Goal: Make many changes at once

Move all NotStarted stories from M-001 to M-002
Change workstream from "eng" to "engineering" for all tasks
Archive all Completed milestones
Update all High priority items to Critical

Dry Run First

Always ask AI to show what will change before executing bulk operations

Advanced Patterns

Spec-Driven Development

sequenceDiagram
    participant You
    participant AI1 as Design AI
    participant Vault
    participant AI2 as Implementation AI

    You->>AI1: Design feature X
    AI1->>Vault: Creates Feature, Stories, Document
    You->>Vault: Reviews on canvas
    You->>AI2: Implement according to DOC-001
    AI2->>Vault: Updates task status as it works
    You->>Vault: Reviews progress

Workflow: 1. Use AI to design and create spec (DOC-xxx) 2. Review spec visually 3. Use different AI to implement based on spec 4. Track progress on canvas

Decision Documentation

Create a decision about [topic]:
- Context: [background]
- Options considered: [A, B, C]
- Decision: [chosen option]
- Rationale: [why]
- Affects: [S-001, M-002]

Benefits: - Historical record of "why" - Links to affected entities - Searchable decision log

Feature Tracking

Create feature "User Notifications":
- Workstream: engineering
- Priority: High
- Implemented by: S-010, S-011, S-012

Show me completion status of F-001
What stories still need to be done for F-001?

Best Practices

Do

  • Use natural language - don't worry about exact syntax
  • Ask AI to explain what it will do before confirming
  • Review changes on canvas after bulk operations
  • Use AI for analysis and insights, not just CRUD
  • Document decisions as you make them

Don't

  • Don't manually edit YAML frontmatter (use AI instead)
  • Don't create complex dependency chains without reviewing
  • Don't skip visual review after major changes
  • Don't forget to sync relationships (AI does this automatically)

Example Conversations

Planning a Feature

You: I need to add user authentication to the app

AI: I can help you plan that. Let me ask a few questions:
    - What authentication methods? (email/password, OAuth, etc.)
    - Do you need password reset functionality?
    - Should this be part of an existing milestone or new one?

You: Email/password and Google OAuth. Yes to password reset. 
     Add to M-001 (MVP Launch)

AI: Great! I'll create:
    - F-001: User Authentication System
    - S-010: Email/Password Authentication (implements F-001)
    - S-011: Google OAuth Integration (implements F-001)
    - S-012: Password Reset Flow (implements F-001)
    All under M-001. Should I proceed?

You: Yes, and add a decision document about why we chose Google over Facebook

AI: ✓ Created F-001, S-010, S-011, S-012
    ✓ Created DEC-005: Choose Google OAuth over Facebook OAuth
    All entities linked appropriately.

Tips for Effective AI Collaboration

Be Specific

❌ "Create some tasks"
✅ "Create 3 tasks for S-001 covering setup, implementation, and testing"

Provide Context

❌ "What's the status?"
✅ "What's the status of M-001 and what's blocking completion?"

Use Iterations

  1. Start broad: "Help me plan a blog feature"
  2. Refine: "Add SEO optimization to that"
  3. Detail: "Break down the SEO story into specific tasks"

Leverage AI's Strengths

  • Analysis: "What's the critical path?"
  • Suggestions: "What am I missing for this feature?"
  • Validation: "Are there any circular dependencies?"
  • Reporting: "Summarize progress this week"

Next Steps