Reference no: EM133873684
Artificial Intelligence Programming in Business Analytics
Assessment - AI-Driven Embeddings for Enhanced Retail Customer Insights
Assignment Overview
In this assignment, you will utilise AI-driven techniques for document retrieval and similarity analysis to derive business insights. You will work with vector databases, LLMs (Large Language Models), and cosine similarity to analyse and recommend improvements for an industry scenario.
You will:
Complete the provided Python code (fill in missing parts).
Answer a multiple-choice questionnaire.
Analyse the results and provide a business recommendation.
Industry Scenario: AI-Powered Market Intelligence for E-commerce
You are working for a data-driven e-commerce company that sells consumer electronics. The company wants to improve customer experience and product recommendations using AI-driven search and analytics.
The company has:
Product descriptions, customer reviews, and support documents in multiple formats (PDF, DOCX, and XLSX).
A search system where users want contextually relevant product recommendations.
A similarity system that helps analyse customer reviews and product descriptions to improve product categorisation.
Your Tasks:
Implement document storage and retrieval using ChromaDB.
Perform cosine similarity analysis to compare customer reviews and product descriptions.
Utilise an LLM (like LLAMA or GPT-Neo) to enhance search results using RAG (Retrieval- Augmented Generation)
Part 1: Complete the Python Code
Complete the Python notebook by filling in the gap.
Create a markdown, write down your name and student number
Part 2: Business Analytics Questionnaire
Based on the results of your AI-powered document retrieval, cosine similarity analysis, and AI- generated recommendations, answer 10 multiple-choice questions.
These questions are designed to test your business analytics thinking and ability to interpret data- driven insights.
The questionnaire will be available Monday 10:00 am W11- Friday 5 pm (AEST) Week 12.
Once students start the questionnaire, they will have 30 minutes to complete 10 multiple-choice questions.
You will enter the "Attempt Questionnaire" on MyKBS in the Assessment Table for A3.
Answer 10 multiple-choice questions.
The questionnaire can only be attempted once.
Backtracking of questions is not allowed. You must complete the question before moving on to the next one. You will not be able to go back to the previous question.
Part 3: Business Analytics Report Task - 1200 words
Your report should be structured as a professional business analytics document intended for senior management and key stakeholders. Ensure a data-driven approach, use tables where necessary, and include actionable insights. Get online assignment help from Ph.D. experts!
Executive Summary (on a single page) - 100 words, 2 marks
Brief overview of the analysis conducted.
Key findings from document retrieval, similarity analysis, and AI-driven recommendations.
Business impact of implementing AI-driven insights.
Business Context and Problem Statement - 100 words, 2 marks
Describe the business challenge that this analysis aims to address.
Explain why enhancing product search, customer sentiment analysis, and AI-powered
recommendations is important for the company.
Use a realistic business scenario, such as:
Low product conversion rates due to ineffective search functionality.
Mismatch between customer expectations and product descriptions.
Poor alignment between marketing messaging and customer sentiment.
Data Analytics Process - 300 words (divided into 3.1 & 3.2), 6 marks
Explain the methodology used to derive business insights from the AI-powered analytics.
Document Retrieval and Vector Search Analysis
Process Explanation: How were documents (product descriptions, customer reviews, and business reports) stored and retrieved using vector search and ChromaDB?
Results Interpretation: What documents were retrieved when querying the database? Provide a table of retrieved documents and explain their relevance.
Cosine Similarity Analysis for Customer Sentiment
Objective: Measuring alignment between customer reviews and product descriptions to detect gaps in product positioning.
Results Presentation: Provide a table of cosine similarity scores between customer reviews and product descriptions.
Insights:
High Similarity (0.8 - 1.0): Product descriptions are aligned with customer expectations.
Medium Similarity (0.5 - 0.8): Some disconnects in customer perception and official messaging.
Low Similarity (0.0 - 0.5): Potential issues with how products are marketed or perceived.
AI-Generated Business Insights (RAG)
How did the AI model (LLAMA) assist in product insights?
Compare AI-generated insights with vector-based search results.
Evaluate if AI-generated responses improved decision-making.
Provide an example of AI-generated business recommendations based on retrieved documents.
Business Recommendations for Stakeholders - 350 words, 8 marks Use diagrams/visual charts to explain your recommendations
Using your findings, propose three to five actionable recommendations for different business functions:
Product Development Team
Improve product descriptions for items with low similarity scores to better align with customer sentiment using AI.
Adjust product messaging based on AI-driven review sentiment insights.
Marketing Strategy
Use AI insights to refine advertising copy for better engagement.
Develop targeted marketing strategies focusing on highly rated product features.
Address pricing concerns identified in customer reviews.
Customer Experience & Support
Implement AI-powered product search enhancements to improve customer experience.
Train support teams to proactively address issues identified in AI-driven insights.
Use AI-driven chatbots to answer customer questions using RAG-based document retrieval.
Business Impact Analysis -100 words, 2 marks
Revenue Impact: How will aligning descriptions and customer sentiment affect conversion rates?
Operational Efficiency: How does using AI reduce manual effort in customer support and marketing?
Competitive Advantage: How does AI-driven retrieval help in gaining an edge over competitors?
Conclusion and Next Steps - 100 words, 2 marks
Summarise how AI-powered analytics enhances business decision-making.
Recommend further AI integration to improve forecasting, demand analysis, and competitor benchmarking.
Suggest future A/B testing of AI-driven content enhancements.