Asha Naik

Data Analyst and Data Scientist, Kotak Mahindra Bank Pvt. Ltd., Bengaluru

April 2023 - March 2024

Conversational-AI

RAG-based document search:

• Built an ingestion pipeline which involves mainly 4 steps: Ingestion, Retrieval, Generation, and Evaluation.
• Loading the document, parsing, and doing the document chunking so that each chunk is stored in the vector database as embeddings, as a node. Retrieve the chunk based on the user's query and generate a response using the given prompt.
• I handled the document ingestion part, used different parsing and chunking strategies, which is a very important and main step in RAG-based document search task.
• Evaluate the process based on similarity metrics by comparing the generated response and ground truth.

Topic Modeler for NPS.

• Designed and implemented a topic modeling system to analyze customer feedback, enhancing the Net Promoter Score (NPS) by improving product and service quality.
• Enabled categorization of customer queries and feedback, helping notify relevant teams to address specific concerns, ultimately leading to improved customer experience and NPS growth.
• Streamlined the customer feedback analysis process, providing actionable insights for improving customer satisfaction across key product and service areas.

Chatbot and WhatsApp Analytics Projects

Chatbot Analytics:

• Monthly Active Users (MAU): 21 lakh users
• Net Promoter Score (NPS): Improved from -3 to 11, significantly enhancing customer satisfaction
• FAQ Performance: Monitored and optimized the performance of frequently asked questions.
• Service Requests (SR) & Enquiries: Tracked and analyzed trends in customer service inquiries
• Credit Card Bill Payments: Analyzed data related to credit card bill payments through the chatbot.
• Recharge Transactions: Assessed recharge transaction data, improving user experience.
• Channel-wise Message Count: Measured message traffic across different channels for performance insights.
• Leads Generated: Tracked and analyzed leads from credit card and loan offers.

WhatsApp Analytics:

• Total Interactions: Tracked customer interactions to date.
• MAU: 7 lakhs monthly active users.
• User-Initiated Messages: Monitored and analyzed trends in user-initiated messages.
• Recharge Payments: Assessed volume and value of successful recharge transactions.
• Top Service Requests: Identified and optimized the most common service requests.
• NPS: Achieved a Net Promoter Score of 8 from 1.

MIS Report:

• Payment and Transaction Data:

Chatbot: Credit Card Payments, Mobile, and DTH Recharge

WhatsApp Banking: Mobile Prepaid and DTH Recharge

• Top 5 Features Usage:

Chatbot: Deliverables, Account Balance Enquiries, Credit Card Enquiries

Transaction History.

WhatsApp: Account Balance Enquiries, Transaction History, Credit Card.

Enquiries, Loan Enquiries.

• Service Requests (SR):

Chatbot: Dispute, Pre-approved Offers, Debit/Credit Card Control, Loan

Applications

WhatsApp: Pre-approved offers, loan offers, credit card statements.

Account and Loan Statements.

OKR (Objectives and Key Results) Report

• Tracked and reported on OKRs related to customer experience and chatbot performance: NPS and MAU.

• Banking Credit Risk Analysis