Business Analyst

# CLAUDE.md — Business Analyst
> Version: 2026-02-17 | Difficulty: beginner

## What Your AI Can Now Do
- Turn a stakeholder conversation into structured user stories with acceptance criteria in minutes
- Map a current-state process and identify bottlenecks automatically from your notes
- Generate a requirements traceability matrix from scattered documents
- Build a gap analysis between current state and target state from two descriptions
- Draft a business case with ROI estimates from your assumptions and data points

---

## Quick Start
1. Create your first project folder: add a brief to `projects/[project-name]/brief.md`
2. Run `/user-stories` on meeting notes to see raw conversation become structured requirements
3. Try `/gap-analysis` with a current-state and target-state description
4. Add Jira and Confluence credentials to `.env` for backlog and wiki integration
5. Run `/process-map` on a workflow description to generate a step-by-step process flow

---

## About Me
I'm a business analyst who bridges the gap between stakeholders and delivery teams. My days involve requirements gathering, process mapping, writing user stories, and making sure what gets built matches what the business actually needs. Clarity and traceability are my currencies.

---

## Pre-Built Commands

| Command | What It Does |
|---------|-------------|
| `/user-stories` | Converts meeting notes into user stories with acceptance criteria |
| `/gap-analysis` | Compares current vs. target state and identifies gaps to close |
| `/process-map` | Generates a step-by-step process flow from a description |
| `/requirements-matrix` | Builds a traceability matrix linking requirements to deliverables |
| `/business-case` | Drafts a business case with costs, benefits, and ROI timeline |

---

## How Claude Connects to Your Tools

| Tool | Why You Need It |
|------|----------------|
| Jira | Create and update user stories directly in your project backlog |
| Confluence | Read existing documentation, publish requirements and process maps |
| Google Drive | Access shared stakeholder docs and save deliverables |
| Slack | Pull context from project channels for requirements clarification |

### Setup
```bash
# Required in .env
JIRA_BASE_URL=https://your-org.atlassian.net
JIRA_API_TOKEN=your-token
JIRA_EMAIL=your-email
CONFLUENCE_BASE_URL=https://your-org.atlassian.net/wiki
```

---

## Analysis Rules
- **MUST:** Never write a user story without understanding the actual end user — stakeholder proxies get it wrong
- **MUST:** Every requirement needs a clear "done" definition — vague acceptance criteria cause rework
- **SHOULD:** Map the current-state process before proposing the future state
- **MUST:** Trace every requirement back to a business objective — orphan requirements waste dev time
- **SHOULD:** Use "As a [user], I want [goal], so that [reason]" format consistently for user stories
- **MUST:** Flag assumptions explicitly — undocumented assumptions become scope creep

---

## Project Structure
```
projects/
  [project-name]/
    brief.md
    requirements/
    process-maps/
    user-stories/
    deliverables/
stakeholders/
templates/
standards/
```

### What These Folders Mean
- `projects/` — One folder per project with all BA artifacts organized by type
- `stakeholders/` — Contact info and context for key stakeholders across projects
- `templates/` — Reusable templates for user stories, BRDs, and process maps
- `standards/` — Naming conventions, definition of done, and quality checklists

---

## Automation

| When | What Happens | If It Fails |
|------|-------------|-------------|
| New file in `requirements/` | Auto-checks for missing acceptance criteria and flags gaps | Review requirements manually against checklist in `standards/` |
| User story created | Validates format and links to parent requirement | Check story format and add traceability link manually |
| Process map updated | Cross-references with requirements for coverage gaps | Run `/requirements-matrix` to check alignment |
| Weekly on Monday | Generates requirements status report across active projects | Run status check on each `projects/` folder individually |

---

## Gotchas
- **MUST:** Talk to actual users, not just their managers — secondhand requirements drift fast
- **MUST:** Version your requirements documents — "final_v3_REAL_final" is not a versioning strategy
- **SHOULD:** Get sign-off on requirements before sprint planning — mid-sprint changes cost 4x
- **MUST:** Document what's out of scope as clearly as what's in scope
- **SHOULD:** Keep a decisions log — "why did we do it this way" saves hours in future sprints

---

## Memory & Learning
Claude builds context over time:
- Stakeholder names, roles, and communication preferences
- Project-specific terminology and business domain language
- Your preferred user story and requirements formats
- Common process patterns in your organization
- Past decisions and their rationale for reference in future discussions

To save something important: just say "remember that [fact]" in conversation.
To review saved memories: use the `/memory` command to open the memory editor.