Meepkun/AI-Support-Ticket-Bot
🤖 Classify support tickets and generate automated responses with this AI-powered bot using a BERT model for efficient customer service.
What's novel
🤖 Classify support tickets and generate automated responses with this AI-powered bot using a BERT model for efficient customer service.
Code Analysis
4 files read · 2 roundsFine-tunes a BERT classifier on 6 hardcoded support tickets for 3 epochs, then classifies new tickets into one of 2 categories and returns a canned text response from a hardcoded dictionary.
Strengths
The BERT classification wrapper is correctly structured and the code is readable and follows a logical flow from data loading through training to inference. The separation of the model class into its own file is a reasonable (if minimal) architectural choice.
Weaknesses
The README is a marketing document that describes features (UI, integrations, downloadable binaries, advanced AI) that do not exist in the code at all. Training BERT on 6 samples is functionally meaningless, there is zero error handling, no tests, and unused dependencies (openai, python-dotenv) are listed in requirements.
Score Breakdown
Signal breakdown
Innovation
Craft
Traction
Scope
Evidence
Commits
5
Contributors
2
Files
7
Active weeks
3
Repository
Language
Python
Stars
1
Forks
0
License
—