CXIntel is a Salesforce-based Customer Experience Intelligence Platform that centralizes customer feedback, applies AI-powered sentiment analysis, and provides real-time insights to agents, managers, and executives.
The application replaces fragmented spreadsheets and manual processes with a secure, automated, and intelligent platform that improves feedback handling, reduces churn, and enhances overall customer experience.
| Feature Implemented | Technology Used |
|---|---|
| Data Model | Custom Object: CustomerFeedback__c, Fields: Feedback_Text__c, Sentiment__c, Sentiment_Score__c, Processed__c |
| Security & Sharing | Experience Cloud Guest Access, Profiles, Permission Sets |
| Business Logic | Validation Rules, Salesforce Flow (case creation), Apex Classes |
| Backend Development | Apex Triggers (auto case creation), Python Scripts (TextBlob/Hugging Face NLP) |
| Frontend Development | Lightning Web Components (Registration, Login, Feedback Form, Dashboard UI) |
| Integration | Salesforce REST API, Python (simple-salesforce), Hugging Face/TextBlob |
| Deployment | GitHub Actions CI/CD, Salesforce DX, VS Code |
| Analytics | Custom Reports & Dashboards (Sentiment Trends, Feedback Volume, Agent Performance) |
| AI Extensions | Einstein Bots (auto-replies) |
- Defined requirements for customer feedback lifecycle.
- Set up Salesforce Experience Cloud site for external customers.
- Configured secure sharing settings for guest and registered users.
- Built
CustomerFeedback__cobject. - Added fields for feedback text, sentiment, score, and processed flag.
- Established lookup relationships for agent assignment.
- Validation Rule → Feedback text must be provided.
- Record-Triggered Flow → Create Case automatically if sentiment is Negative.
- Email through Flows → Sends E-mail to the user who have submitted the feedback
- Developed Python script with TextBlob/Hugging Face for sentiment analysis.
- Used
simple-salesforcefor CRUD operations on Salesforce records. - Automated updates for sentiment score & classification.
- Scheduled with Windows Task Scheduler / GitHub Actions.
- Custom Registration & Login LWCs integrated with
Registration__cobject. - Implemented password hashing (SHA-256) for secure storage.
- Enforced password rules: 10+ chars, uppercase, number, special character.
- Built Feedback Portal with Lightning Web Components.
- Added Registration ↔ Login redirects.
- Created UI for feedback submission, FAQs, and discussion forums.
- Integrated Einstein Bots for automated responses.
- Experimented with GPT-powered suggestions for agent replies.
- Dashboards created for:
- Sentiment distribution (Positive, Neutral, Negative).
- Feedback volume trends.
- Agent resolution performance.
- Reports for executives to monitor CX health and prevent churn.
- Full demo flow:
- Customer registers → logs in → submits feedback.
- Python script analyzes feedback → updates Salesforce.
- Positive feedback → collects testimonial
- Neutral feedback → collects survey
- Negative feedback → auto case creation for agent.
- Dashboard updates in real-time.
- Demonstrated business impact: faster resolutions, AI-driven insights, improved customer satisfaction.
- Prevent Missed Feedback: Automated processing and sentiment analysis.
- Streamline Operations: Centralized customer feedback system.
- Enhance Compliance & Security: Encrypted authentication & secure guest access.
- AI-Powered Insights: Predict churn, identify customer pain points, improve service quality.
- Salesforce (Experience Cloud, Apex, LWC, Flows, Reports & Dashboards)
- Python (TextBlob, Hugging Face Transformers, simple-salesforce)
- Deployment Tools: GitHub Actions, VS Code, Salesforce DX
- AI Extensions: Einstein Bots
