Phi-2 Multi-Issue Transcript Analysis Model

This model is based on Microsoft's Phi-2 for analyzing customer service transcripts with multiple issues. It can:

Identify primary and secondary issues
Analyze customer sentiment
Rate agent performance
Track resolution status
Predict CSAT scores
Extract key actions and outcomes

Model Details

Base Model: microsoft/phi-2
Task: Multi-issue customer service transcript analysis
Training Data: Customer service transcripts with multiple issues
Output Format: Structured JSON with detailed analysis

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

# Load model and tokenizer
model = AutoModelForCausalLM.from_pretrained("chendren/phi2-multi-issue-analysis")
tokenizer = AutoTokenizer.from_pretrained("chendren/phi2-multi-issue-analysis")

# Prepare input
transcript = """[Your customer service transcript here]"""

# Generate analysis
inputs = tokenizer(transcript, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=512)
analysis = tokenizer.decode(outputs[0])

Example Output

{
  "primary_issue": "Internet connection drops",
  "secondary_issues": [
    "Signal interference",
    "Router firmware outdated"
  ],
  "customer_sentiment": "negative",
  "agent_performance": {
    "rating": 4,
    "justification": "Agent was helpful and provided clear instructions"
  },
  "resolution_status": "resolved",
  "follow_up_needed": false,
  "key_points": [
    "Customer experienced internet drops",
    "Agent guided through troubleshooting",
    "Issue resolved with firmware update"
  ],
  "issues": [
    "Intermittent connection drops",
    "WiFi interference",
    "Outdated firmware"
  ],
  "actions": [
    "Diagnosed signal fluctuations",
    "Updated router firmware",
    "Provided monitoring instructions"
  ],
  "outcomes": [
    "Connection stability improved",
    "Firmware updated successfully"
  ],
  "predicted_csat": 4
}

Limitations

Designed specifically for customer service transcripts
Best performance with clear dialogue format
May require adjustment for different transcript formats

Citation

If you use this model, please cite:

@misc{phi2-multi-issue-analysis,
  author = {args.username},
  title = {Phi-2 Multi-Issue Transcript Analysis Model},
  year = {2025},
  publisher = {Hugging Face},
  journal = {Hugging Face Model Hub},
  howpublished = {https://huggingface.co/chendren/phi2-multi-issue-analysis}
}