FCA Vulnerability Analyzer

AI-powered tool for analyzing financial advisor-client conversations to identify vulnerability indicators per FCA guidelines.

Features

• LangChain Integration: Uses GPT-5 with structured prompt templates
• FCA Compliance: Identifies vulnerabilities across 4 key drivers (Health, Life Events, Resilience, Capability)
• JSON Processing: Handles conversation transcripts in JSON format
• Structured Output: Returns detailed vulnerability analysis with advisor recommendations

Setup

1. Install dependencies:

pip install -r requirements.txt

2. Set OpenAI API key in .env:

OPENAI_API_KEY=your_api_key_here

Usage

1. Place transcript JSON files in inputs/ directory
2. Update main.py to reference your input file
3. Run analysis:

python main.py

Results are saved to outputs/analysis_results.json

Project Structure

fca/
├── inputs/           # Transcript JSON files
├── outputs/          # Analysis results
├── fca_analyzer.py   # Core analyzer with LangChain
├── main.py          # Entry point
└── requirements.txt # Dependencies

Output Format

Each vulnerability finding includes:

• client_speech: Exact quote indicating vulnerability
• vulnerability_classification: FCA category (Health/Life Events/Resilience/Capability)
• reason: Explanation of vulnerability indicator
• advisor_actions: Recommended actions for advisor