docs/QUICK_START.md

Quick Start: Fraud Assistant with Confluence Integration

Follow these steps to add Confluence search to your fraud explainability assistant in under 10 minutes.

---

TL;DR (For the Impatient)

# 1. Install dependencies
pip install -r requirements-with-confluence.txt

# 2. Create .env file with your Confluence credentials
cp .env.example .env
nano .env  # Add your Confluence URL, email, and API token

# 3. Test it
python confluence_integration_example.py

Then ask: "What does our fair lending policy say about synthetic ID detection?"

---

Step-by-Step Setup

Step 1: Get Confluence API Token (2 minutes)

1. Go to https://id.atlassian.com/manage-profile/security/api-tokens
2. Click "Create API token"
3. Label it: "Fraud Assistant"
4. Copy the token (you won't see it again!)

Step 2: Configure Environment (1 minute)

cd fraud_model_explainability_assistant

# Copy example file
cp .env.example .env

# Edit with your credentials
nano .env  # or use your preferred editor

Edit these 3 lines in .env:

CONFLUENCE_URL=https://yourcompany.atlassian.net  # YOUR Confluence URL
CONFLUENCE_EMAIL=your.email@company.com           # YOUR email
CONFLUENCE_API_TOKEN=ATATT3xFfGF0...              # YOUR token (paste here)

Save and close.

Step 3: Install Dependencies (2 minutes)

# Install everything needed
pip install -r requirements-with-confluence.txt

# Verify installation
python -c "from confluence_ingestor import ConfluenceRAG; print('✅ Ready!')"

Step 4: Identify Confluence Spaces (1 minute)

Find the "space keys" for your Confluence spaces:

1. Open your Confluence
2. Go to a space
3. Look at the URL: https://yourcompany.atlassian.net/wiki/spaces/SPACEKEY/
4. The SPACEKEY is what you need

Recommended spaces:

• Model governance docs (e.g., fraud-model-governance)
• Compliance policies (e.g., compliance-policies)
• Investigation playbooks (e.g., fraud-investigation)

Step 5: Update Space Keys in Code (1 minute)

Edit confluence_integration_example.py line 88:

spaces = {
    "your-governance-space": 100,   # Replace with YOUR space key
    "your-compliance-space": 50,    # Replace with YOUR space key
    "your-playbooks-space": 75,     # Replace with YOUR space key
}

Step 6: Run Initial Ingestion (5 minutes)

This downloads and indexes your Confluence content (one-time):

python confluence_integration_example.py

What happens:

1. Connects to Confluence ✅
2. Downloads pages from your spaces (~5 min)
3. Creates searchable vector database
4. Launches web UI

First time only: Downloads ~400MB ML model (free, runs locally)

Step 7: Test It! (1 minute)

1. Open http://localhost:7860 in your browser
2. Try these questions:

Question 1 (tests Confluence search):

What does our fair lending policy say about phone type features?

Question 2 (tests fraud analysis):

Why was application APP-78432 flagged as high risk?

Question 3 (tests compliance with policy citation):

Check fair lending compliance for APP-55555 and cite relevant policies

If the agent cites specific Confluence pages in its response, it's working! 🎉

---

What You Just Built

Your fraud assistant now has access to:

• ✅ Company fair lending policies
• ✅ Model governance documentation
• ✅ Investigation procedures
• ✅ Compliance guidelines
• ✅ Historical precedents

Use cases unlocked:

• Executive briefings with policy citations
• Regulatory examination support
• Analyst training and investigation support
• Audit trail documentation

---

Troubleshooting

"3 validation errors for ConfluenceConfig"

Problem: Missing Confluence credentials

Fix:

# Check if .env exists
ls -la .env

# If missing:
cp .env.example .env

# Edit with your credentials
nano .env

# Verify:
cat .env | grep CONFLUENCE

"401 Unauthorized"

Problem: Wrong email or API token

Fix:

1. Re-generate API token at https://id.atlassian.com/manage-profile/security/api-tokens
2. Double-check email is correct
3. Remove any extra spaces in .env

"Space 'xyz' not found"

Problem: Space key doesn't exist or you don't have access

Fix:

1. Check the space key in Confluence URL
2. Verify you can view the space when logged in
3. Update space keys in confluence_integration_example.py line 88

"Module not found: confluence_ingestor"

Problem: Package not installed

Fix:

pip install -r requirements-with-confluence.txt

---

Next Steps

Once it's working:

1. Add more spaces: Edit the spaces dict in the code
2. Customize prompts: Update ENHANCED_PROMPT for your use case
3. Set up auto-updates: See CONFLUENCE_INTEGRATION_ANALYSIS.md
4. Share with team: Deploy on internal server
5. Monitor usage: Track which policies are searched most

---

Full Documentation

• Detailed Setup: See CONFLUENCE_SETUP_GUIDE.md
• Architecture & ROI: See CONFLUENCE_INTEGRATION_ANALYSIS.md
• Confluence-Ingestor Docs: See ../confluence-ingestor/README.md

---

Summary Checklist

• Get Confluence API token (2 min)
• Create .env with credentials (1 min)
• Install dependencies (2 min)
• Find your Confluence space keys (1 min)
• Update space keys in code (1 min)
• Run initial ingestion (5 min)
• Test with sample questions (1 min)

Total Time: ~10 minutes

---

Support

Having issues? Check:

1. CONFLUENCE_SETUP_GUIDE.md - Detailed troubleshooting
2. CONFLUENCE_INTEGRATION_ANALYSIS.md - Architecture details
3. The "Troubleshooting" section above

Still stuck? Review the error message - it usually tells you what's missing!

---

You're done! Your fraud assistant can now search company documentation and cite specific policies. 🚀