Vision & Planning Agents¶
This document provides a comprehensive overview of Vision & Planning agents in the ConnectSoft AI Software Factory. It is written for product managers, architects, and stakeholders who need to understand how the Factory transforms product ideas into actionable plans.
Vision & Planning agents operate early in the development lifecycle, transforming vague ideas into defined product visions, scopes, epics, and feature sets. They ensure alignment with business goals and technical feasibility before detailed architecture and coding begin.
Note
Vision & Planning agents work with product managers and stakeholders to refine requirements. They don't generate final code—they create the foundation that Architect and Engineering agents build upon.
Agent Cluster Composition¶
The Vision and Product Planning cluster consists of 4 specialized agents that work together to transform unstructured business ideas into structured, execution-ready product plans:
| Agent | Core Function | Primary Output |
|---|---|---|
| Vision Architect Agent | Converts business ideas into structured software vision documents | Vision Document (vision.md, vision.json) |
| Product Manager Agent | Builds product roadmap, MVP plan, personas, and release phases | Product Plan (product-plan.md, features.json) |
| Product Owner Agent | Translates plan into backlog-ready user stories and acceptance criteria | User Stories (user-stories.md, acceptance-criteria.md) |
| Business Analyst Agent | Extracts structured business requirements, rules, and processes | Business Requirements (requirements.md, rules.md) |
Each agent builds upon the outputs of previous agents, forming a composable, semantic planning pipeline.
Mission and Scope¶
Vision & Planning agents are responsible for:
- Transforming Ideas into Visions - Converting high-level concepts into clear product visions
- Scoping and Slicing - Breaking down large ideas into manageable epics and features
- Requirements Refinement - Clarifying ambiguous requirements into actionable specifications
- Risk and Complexity Assessment - Identifying risks and complexity early in the process
- Business Rule Extraction - Formalizing business rules and processes
- Product Roadmap Creation - Planning phased delivery and MVP definition
What They Do:
- Refine product ideas into detailed requirements
- Create epics, user stories, and feature breakdowns
- Align requirements with business goals
- Assess technical feasibility and risks
- Create product roadmaps and release plans
- Extract and formalize business rules
- Define personas and user needs
What They Do NOT Do:
- Generate final code or architecture
- Make detailed technical decisions
- Design bounded contexts or APIs
- Implement features
Inputs and Outputs¶
Inputs¶
| Input | Source | Description |
|---|---|---|
| High-Level Business Goals | Product managers, stakeholders | What problem are we solving? |
| Problem Statements | Business requirements | User needs and pain points |
| Constraints | Business context | Budget, timelines, tech stack preferences |
| Existing System Context | Knowledge system | Current systems and integrations |
| Market Research | External sources | Competitive analysis, market trends |
Outputs¶
| Output | Consumer | Description |
|---|---|---|
| Product/Feature Briefs | Architect Agents | High-level product vision and goals |
| Epic & Feature Outlines | Architect Agents, Engineering Agents | Detailed feature breakdowns |
| User Stories | Engineering Agents | Actionable user stories with acceptance criteria |
| Product Roadmaps | Product managers, stakeholders | Phased delivery plans |
| Risk & Complexity Tags | Architect Agents | Early risk identification |
Core Capabilities¶
Market and Solution Framing¶
- Problem-Solution Fit - Analyze if the solution fits the problem
- Market Positioning - Understand competitive landscape
- Value Proposition - Define clear value proposition
- User Personas - Identify target users and their needs
High-Level Scoping and Slicing¶
- Epic Decomposition - Break large features into epics
- Feature Prioritization - Prioritize features by business value
- MVP Definition - Identify minimum viable product scope
- Release Planning - Plan phased releases
Risk and Complexity Tagging¶
- Technical Risk Assessment - Identify technical risks early
- Complexity Estimation - Estimate feature complexity
- Dependency Identification - Identify dependencies between features
- Feasibility Analysis - Assess if features are feasible
Mapping to Bounded Contexts¶
- Domain Boundary Identification - Suggest initial bounded context boundaries
- Context Relationships - Identify relationships between contexts
- Integration Points - Identify integration needs
Typical Workflows¶
Workflow: New SaaS Product Idea¶
- Receive Input - Product manager provides high-level idea
- Refine Vision - Vision agent refines idea into product vision
- Define Scope - Planning agent breaks down into epics and features
- Assess Risks - Identify risks and complexity
- Create Roadmap - Plan phased delivery
- Hand Off to Architect - Pass refined requirements to Architect agents
Workflow: Feature Addition to Existing Product¶
- Receive Feature Request - New feature request for existing product
- Analyze Context - Review existing system and domain
- Refine Requirements - Clarify feature requirements
- Break Down - Create user stories and acceptance criteria
- Assess Impact - Identify impact on existing features
- Plan Integration - Plan integration with existing system
Agent Collaboration Flow¶
The Vision and Planning agents collaborate through event-driven handoffs:
sequenceDiagram
participant Human as Human Input
participant VA as Vision Architect
participant PM as Product Manager
participant PO as Product Owner
participant BA as Business Analyst
participant Arch as Architect Agents
Human->>VA: Business Idea/Prompt
VA->>VA: Create Vision Document
VA->>PM: Emit VisionDocumentCreated
VA->>BA: Emit VisionDocumentCreated (parallel)
PM->>PM: Create Product Plan
BA->>BA: Extract Requirements
PM->>PO: Emit ProductPlanCreated
BA->>PO: Emit BusinessRequirementsReady
PO->>PO: Generate User Stories
PO->>Arch: Emit BacklogReady
Standard Sequential Flow¶
flowchart LR
A[Product Idea] --> B[Vision Architect Agent]
B -->|VisionDocumentCreated| C[Product Manager Agent]
C -->|ProductPlanCreated| D[Product Owner Agent]
D -->|BacklogReady| E[Business Analyst Agent]
E -->|BusinessRequirementsReady| F[Architect Agents]
F --> G[Architecture Blueprint]
style A fill:#e1f5ff
style B fill:#fff4e1
style C fill:#e8f5e9
style D fill:#f3e5f5
style E fill:#fce4ec
style G fill:#e0f2f1
Alternative: Fork-Join Flow¶
For complex projects, Product Manager and Business Analyst can work in parallel:
flowchart TD
Vision[Vision Architect Agent] -->|VisionDocumentCreated| PM[Product Manager Agent]
Vision -->|VisionDocumentCreated| BA[Business Analyst Agent]
PM -->|ProductPlanCreated| PO[Product Owner Agent]
BA -->|BusinessRequirementsReady| PO
PO -->|BacklogReady| Arch[Architect Agents]
Individual Agent Details¶
Vision Architect Agent¶
Role: Translates freeform input into structured Vision Document
Inputs: - Human idea, objective prompt, constraints - Use case context, project metadata - Optional: Template info, runtime constraints
Outputs:
- vision.md - Human-readable vision document
- vision.json - Machine-consumable structured version
- Metadata with traceId, software_type, confidence_score
- Event: VisionDocumentCreated
Key Responsibilities: - Parse freeform input to extract problem, goals, constraints - Formalize business problem definition - Identify strategic angles and market fit - Draft high-level solution approach - Classify software type (SaaS, mobile app, API, etc.) - Suggest initial feature mapping - Define personas (users, admins, stakeholders) - Formulate success metrics - Generate structured vision artifacts
See: Detailed Vision Architect Agent specification in Factory Documentation
Product Manager Agent¶
Role: Plans product roadmap, MVP cutline, and release phases
Inputs: - Vision Document from Vision Architect - Business Requirements (if Business Analyst ran early) - Strategic goals, personas - Market research and competitive analysis
Outputs:
- product-plan.md - Product roadmap and strategy
- features.json - Feature breakdown
- epics.yaml - Epic definitions
- MVP scope definition
- Release phases and milestones
- Event: ProductPlanCreated
Key Responsibilities: - Build product roadmap with phased delivery - Define MVP scope and cutline - Prioritize features by business value - Plan release phases and milestones - Align product strategy with business goals - Create feature breakdowns and epics - Define success metrics and KPIs
See: Detailed Product Manager Agent specification in Factory Documentation
Product Owner Agent¶
Role: Generates backlog-ready user stories and acceptance criteria
Inputs: - Product Plan from Product Manager - Business Requirements from Business Analyst - Personas and user needs - Constraints and technical limitations
Outputs:
- user-stories.md - Detailed user stories
- acceptance-criteria.md - Acceptance criteria for each story
- story-map.json - Story mapping and prioritization
- Backlog ready for development
- Event: BacklogReady
Key Responsibilities: - Translate product plan into actionable user stories - Define acceptance criteria for each story - Create story maps and prioritization - Ensure stories are INVEST-compliant (Independent, Negotiable, Valuable, Estimable, Small, Testable) - Link stories to epics and features - Prepare backlog for Architect and Engineering agents
See: Detailed Product Owner Agent specification in Factory Documentation
Business Analyst Agent¶
Role: Extracts structured business requirements, rules, and processes
Inputs: - Vision Document from Vision Architect - Product Plan (if available) - Domain constraints and regulations - Existing business processes
Outputs:
- requirements.md - Functional requirements
- rules.md - Business rules and policies
- bpmn.mmd - Business process models
- gap-analysis.md - Gap analysis
- Event: BusinessRequirementsReady
Key Responsibilities: - Extract functional requirements from vision - Formalize business rules and policies - Model business processes (BPMN) - Identify gaps and dependencies - Ensure compliance with domain regulations - Validate requirements completeness - Enrich backlog items with business context
Positioning Flexibility: - After Product Owner - Validates and enriches backlog (standard flow) - Before Product Manager - Early requirements decomposition (complex domains) - Parallel with Product Manager - Fork-join model for speed
See: Detailed Business Analyst Agent specification in Factory Documentation
Collaboration with Other Agents¶
With Architect Agents¶
Hand Off:
- Product briefs and feature outlines
- Initial bounded context suggestions
- Risk and complexity assessments
Receive Feedback:
- Feasibility concerns
- Architecture constraints
- Technical limitations
With Engineering Agents¶
Hand Off:
- User stories with acceptance criteria
- Feature specifications
- Business rules and requirements
Receive Feedback:
- Implementation complexity
- Technical blockers
- Estimation adjustments
With QA Agents¶
Receive Feedback:
- Testability concerns
- Quality risks
- Edge case identification
Example Prompts and Tasks¶
Example 1: New Product Vision¶
Task: "Given this high-level requirement: 'We need a subscription billing system for our SaaS product', draft epics and key features."
Output:
- Epic: Subscription Management
- Epic: Invoice Generation
- Epic: Payment Processing
- Epic: Billing Analytics
Example 2: MVP Scoping¶
Task: "Refine this backlog to fit a 3-month MVP timeline."
Input: Large feature backlog Output: Prioritized MVP scope with phased delivery plan
Example 3: Feature Refinement¶
Task: "Clarify this ambiguous requirement: 'Users should be able to manage their subscriptions'."
Output: Detailed user stories with acceptance criteria:
- As a tenant admin, I want to create subscriptions
- As a tenant admin, I want to cancel subscriptions
- As a tenant admin, I want to upgrade/downgrade subscriptions
Example 4: Risk Assessment¶
Task: "Assess risks and complexity for this feature: 'Real-time payment processing with Stripe'."
Output:
- Risk: Payment provider downtime
- Risk: PCI compliance requirements
- Complexity: High (external integration, error handling)
Example 5: Release Planning¶
Task: "Create a 6-month release plan for this product roadmap."
Output: Phased release plan with milestones and deliverables
KPIs and Evaluation Criteria¶
| KPI | Description | How We Measure |
|---|---|---|
| Clarity of Epics/Requirements | How clear and actionable are the outputs | Number of clarification requests from downstream agents |
| Alignment with Business Goals | How well outputs align with business objectives | Stakeholder review and approval rate |
| Reduction in Rework | How much rework is needed downstream | Changes requested by Architect/Engineering agents |
| Time to First Blueprint | Speed from idea to architecture blueprint | Time from Vision input to Architect output |
| Feature Completeness | How complete are the feature specifications | Coverage of acceptance criteria and edge cases |
Tip
Good Vision & Planning outputs reduce downstream rework significantly. Clear, well-scoped requirements lead to faster architecture and implementation.
Risks and Anti-Patterns¶
Risk 1: Overdesigning Future Phases¶
Problem: Planning agents focus too much on future phases instead of MVP.
Solution: Enforce MVP-first thinking. Prioritize immediate needs over future features.
Risk 2: Conflicting Goals or Ambiguous Scope¶
Problem: Requirements are ambiguous or contain conflicting goals.
Solution: Require clear acceptance criteria. Flag ambiguities for human review.
Risk 3: Ignoring Technical Constraints¶
Problem: Planning agents ignore technical constraints or feasibility.
Solution: Include technical feasibility checks. Get Architect agent feedback early.
Risk 4: Scope Creep¶
Problem: Scope expands beyond original intent.
Solution: Enforce scope boundaries. Require explicit approval for scope changes.
Warning
Biggest Risk: Overdesigning future phases instead of focusing on MVP. Vision & Planning agents should prioritize immediate value delivery over comprehensive future planning. Start small, iterate, and learn.
Event-Driven Activation¶
All Vision and Planning agents communicate via structured events:
| Event | Emitted By | Consumed By | Description |
|---|---|---|---|
VisionDocumentCreated |
Vision Architect Agent | Product Manager, Business Analyst | Vision document ready for planning |
ProductPlanCreated |
Product Manager Agent | Product Owner | Product plan ready for backlog creation |
BusinessRequirementsReady |
Business Analyst Agent | Product Owner, Product Manager | Business requirements ready |
BacklogReady |
Product Owner Agent | Architect Agents, Engineering Agents | Backlog ready for architecture and development |
Related Documents¶
- Agent System Overview - How agents work together
- Agent Execution Flow - Execution flow details
- Agent Collaboration Patterns - Patterns for agent interactions
- Research and UX/UI Design Agents - Next step in the pipeline
- Architect Agents - Architecture design phase
- Modularization - Bounded context guidance
- AI-First Development - AI-first principles