Add example: financial-advice-bot evaluated against fair-lending + EU AI Act

[kmadan]

Goal

Create examples/financial-advice-bot/ following the six-file shape, demonstrating an LLM that gives general financial education (never specific advice) evaluated against gopal's industry_specific/bfs/v1/ fair-lending + model-risk policies.

Why this matters

Financial-services AI is a high-touch regulated space (MiFID II, FCA, SEC, MAS notices). A documented example accelerates adoption in fintech and bank IT.

Files likely involved

• New: examples/financial-advice-bot/ (six-file shape — see examples/customer-support-bot/ for the pattern)
• examples/README.md table entry

Acceptance criteria

• Contract with 6-10 interactions spanning factual questions, advice refusals, protected-attribute neutrality, and complaint routing
• policy_config.yaml selecting eu_ai_act + bfs + global
• run.py is a verbatim copy of the customer-support-bot pattern adapted for finance
• expected_report.md covers the pass case AND common failure modes (model emits specific advice; conditions on demographic info)
• Honest scope: a green report does not authorise financial-advice delivery — say so explicitly

Help

Comment here if you want to work on this and I will help you get started. First-time contributors are welcome — open a draft PR early and I'll review.