An Agentic AI-powered multi-agent system designed to automate the analysis, classification, prioritization, and ticket generation workflow for SaaS and mobile application feedback.
The platform processes app store reviews and customer support emails, extracts actionable insights, identifies bugs and feature requests, assigns priorities, and generates structured engineering tickets automatically. Built using CrewAI/AutoGen and Streamlit, the system improves operational efficiency, reduces manual triaging effort, and ensures consistency across support and engineering workflows.
- Multi-Agent AI Architecture
- Automated CSV Feedback Processing
- NLP-Based Feedback Classification
- Bug Severity & Technical Detail Extraction
- Feature Request Analysis
- Intelligent Priority Assignment
- Structured Ticket Generation
- Quality Validation & Consistency Checks
- Streamlit Monitoring Dashboard
- Processing Logs & Metrics Tracking
- Configurable Rules & Thresholds
- Robust Error Handling
-
User feedback is collected from CSV files
-
CSV Reader Agent ingests reviews and support emails
-
Feedback Classifier Agent categorizes feedback into:
- Bug
- Feature Request
- Praise
- Complaint
- Spam
-
Bug Analysis Agent extracts severity and technical details
-
Feature Extractor Agent analyzes feature requests
-
Ticket Creator Agent generates structured tickets
-
Quality Critic Agent validates ticket consistency and completeness
-
Final tickets, logs, and metrics are stored for monitoring and analysis
- Streamlit
- CrewAI / AutoGen
- Python 3.11
- Pandas
- OpenAI / LLM APIs
- CSV-Based Processing Pipeline
agentic-feedback-system/
│
├── agents/
│ ├── bug_analysis_agent.py
│ ├── classifier_agent.py
│ ├── csv_reader_agent.py
│ ├── feature_extractor_agent.py
│ ├── quality_critic_agent.py
│ └── ticket_creator_agent.py
│
├── data/
│ ├── app_store_reviews.csv
│ ├── support_emails.csv
│ └── expected_classifications.csv
│
├── output/
│ ├── generated_tickets.csv
│ ├── processing_log.csv
│ └── metrics.csv
│
├── utils/
│ ├── classifier.py
│ ├── logger.py
│ └── metrics.py
│
├── app.py
├── main.py
├── requirements.txt
└── .gitignoregit clone https://github.com/pathak-builds/agentic-feedback-system.git
cd agentic-feedback-systempy -3.11 -m venv venvActivate the environment:
venv\Scripts\activatesource venv/bin/activatepip install -r requirements.txtpython main.pystreamlit run app.pyThe system processes:
app_store_reviews.csvsupport_emails.csvexpected_classifications.csv
Generated outputs include:
generated_tickets.csvprocessing_log.csvmetrics.csv
- Detect app crashes and login failures
- Identify high-priority bugs automatically
- Extract feature requests from customer feedback
- Prioritize engineering tickets intelligently
- Monitor feedback analytics through dashboard
- Reduce manual customer support triaging
- Jira / Trello / Slack Integration
- Real-Time Feedback Streaming APIs
- Advanced Sentiment & Emotion Analysis
- Vector Database Memory Integration
- RAG-Based Contextual Ticket Enrichment
- Multi-Language Feedback Support
- Human-in-the-Loop Approval Workflow
- Automated Root Cause Analysis
- Cloud Deployment with CI/CD
- AI-Powered Trend Forecasting Dashboard
Nimish Pathak