Documenting Business Processes

Documenting Business Processes

Document the business processes and workflows a system implements: how domain entities move through the system, what triggers each step, what the outcomes are, and what can go wrong. Business process docs tell the end-to-end story that no single component's documentation can tell on its own.

Core Principles

1. End-to-End Perspective: Process docs describe a complete flow from trigger to outcome. They cross component boundaries — that's their purpose.
2. Business Language: Describe processes in business terms, referencing domain concepts. "User submits registration form" not "POST /api/v1/users handler validates input."
3. NOT Development Processes: This skill documents business workflows the system implements (user registration, order fulfillment). Development processes (how to test, deploy, contribute) belong in skills and contribution guidelines — not here.
4. No Duplication Across Levels: Reference domain docs for entity definitions and architecture docs for system structure. Process docs describe the FLOW — how entities traverse the architecture. Don't redefine domain terms or re-explain architectural patterns.
5. Include Failure Paths: Happy paths are easy. Document what happens when things go wrong — failed payments, validation errors, timeouts, partial completions.

Documentation Hierarchy

Domain              ← Defines the entities involved in processes
  ↓
Architecture        ← Defines the system structure processes traverse
  ↓
Business Processes  ← You are here (end-to-end flows across the system)
  ↓
Components          ← Implements individual steps of processes

Workflow

Step 1: Discover Existing Documentation

Before writing anything:

1. Check for existing process docs, flow diagrams, or sequence diagrams
2. Read AGENTS.md for pointers to existing documentation
3. Check domain and architecture docs for references to processes
4. If process docs already exist: Understand their structure and extend them

Step 2: Trace the Process

Follow the targeted process through the codebase:

1. Identify the trigger — what initiates this process (user action, scheduled job, external event)
2. Trace the happy path step by step across component boundaries
3. Identify decision points — where the flow branches based on conditions
4. Map failure modes — what can go wrong at each step and how the system handles it
5. Identify the outcomes — what state changes when the process completes (or fails)
6. Note any asynchronous steps, retries, or eventual consistency patterns

Stay within the requested scope. Document one process at a time.

Step 3: Write Process Documentation

Follow the project's existing doc style. If none exists, use this structure:

# <Process Name>

## Overview
<What this process accomplishes from a business perspective — 2-3 sentences>

## Trigger
<What initiates this process — user action, system event, schedule, external call>

## Actors
<Who or what is involved — users, services, external systems>
- <Actor>: <role in this process>

## Diagram
<Mermaid diagram visualizing the process flow — see diagram guidance below>

## Flow

### Happy Path
1. <Step 1> — <what happens, which component handles it>
2. <Step 2> — <what happens, which component handles it>
3. ...
<Result: what state changes when the process completes successfully>

### Failure Scenarios

#### <Failure Scenario 1>
- **Trigger**: <What causes this failure>
- **At step**: <Where in the flow it occurs>
- **Handling**: <How the system responds — rollback, retry, partial completion, error state>
- **User impact**: <What the user experiences>

#### <Failure Scenario 2>
...

## State Changes
<What entities are created, modified, or deleted when this process completes>
- <Entity>: <from state> → <to state>

## Dependencies
<External systems, services, or conditions this process relies on>

Mermaid Diagrams

Every process doc should include a mermaid diagram placed before the textual flow description. The diagram gives humans an instant visual overview; the text provides the detail.

Choose the diagram type based on the process shape:

• Flowchart (flowchart TD): Best for most processes — decision branches, parallel paths, error terminals. Use subgraphs to group related phases.
• Sequence diagram (sequenceDiagram): Best when the process is a back-and-forth between distinct actors or systems (e.g., sourcing chain, API handshakes).

Guidelines:

• Keep diagrams focused — show the main flow and key decision points, not every edge case
• Use descriptive node labels in business language, not function names
• Mark error terminals distinctly (e.g., red styling or stop symbols)
• Use subgraphs to separate phases when a process has distinct stages

Sub-Process Decomposition

When tracing a process, some steps may be complex enough to warrant their own dedicated process doc. Decompose a step into a sub-process when:

• The step has its own decision branches, failure modes, or actors beyond the parent flow
• The step involves multiple sequential actions that would bloat the parent doc
• The step is reusable — referenced by multiple parent processes
• Documenting it inline would make the parent doc's flow section harder to scan

When you identify a sub-process:

1. In the parent doc: Replace the inline description with a link to the sub-process doc. Keep only a one-line summary in the flow step — the detail lives in the sub-process doc.

   3. **[Check system compatibility][compat-check]** — Verify the system meets minimum requirements
   

2. Add a Sub-Processes table to the parent doc (if one doesn't exist) listing all sub-processes with links and brief descriptions.
3. Create the sub-process doc using the same template as the parent — Overview, Trigger, Actors, Diagram, Flow, Failure Scenarios, State Changes, Dependencies. The trigger should reference the parent process (e.g., "Called during step 3 of [installation][installation]").
4. Match existing patterns — if the parent doc already has sub-process links, follow the exact same structure (link style, summary length, Sub-Processes table format). Consistency across sub-processes matters.

When NOT to decompose:

• The step is a single action with no branching or failure modes
• The step's behavior is fully captured in one sentence
• Breaking it out would create a trivially short doc that adds navigation overhead without value

Step 4: Update AGENTS.md Pointers

If new documentation files were created, propose adding a pointer in AGENTS.md:

## Business Processes
See `docs/processes/<process>.md` for <brief description>.

Never put process details into AGENTS.md itself — only add a pointer.

Integration with Other Skills

• documenting-domain: Process docs reference domain entities (e.g., "creates a new User — see domain/users.md for the User entity definition"). Never redefine domain terms here.
• documenting-architecture: Process docs reference architectural patterns when relevant (e.g., "this step publishes an event via the event bus — see architecture/event-bus.md"). Never re-explain the architecture here.
• documenting-components: Component docs describe how individual steps are implemented; process docs describe the end-to-end flow across components.
• project-feature-planning: Process docs help agents understand the full impact of changes — modifying one step of a process affects the entire flow.

Rules

• Never document development processes — testing, deployment, and CI/CD belong in skills and contribution guidelines
• Never document all processes at once — only the targeted process
• Never redefine domain concepts or architectural patterns — reference the appropriate docs
• Always include failure paths — happy-path-only docs are incomplete and misleading
• Business language first — describe what happens from the business perspective, then note which components are involved
• Use reference-style links — when linking to other docs or source files, use reference links ([text][ref] with [ref]: path at the bottom of the file) rather than inline links. They read better in source and are easier to maintain.
• Decompose complex steps into sub-processes — if a step has its own decision branches, failure modes, or multiple sequential actions, it needs its own doc. Don't inline what should be a sub-process.
• Match existing sub-process patterns — if a parent doc already has sub-process links, new sub-processes must follow the exact same structure and link style
• Propose structure first — if no process docs exist yet, propose a directory structure and format before creating files