Biswas Khatiwada | Iowa State University | AI 2010

AI-Powered Customer Review Sentiment Analyzer

Using Natural Language Processing to turn 75 customer reviews into actionable marketing intelligence for small businesses — no coding required.

49% Positive Reviews
40% Negative Reviews
3.47★ Avg Rating
75 Reviews Analyzed

The Problem We Solved

The Problem

Small restaurants like Bella's Bistro receive hundreds of customer reviews on Google and Yelp every month. Owners and managers simply don't have the time to read every review individually. Critical feedback about food quality, wait times, pricing, and service gets buried and goes unread. As a result, important patterns — like a spike in complaints about slow service during the holiday season — are never identified. This leads to uninformed marketing decisions, unresolved customer pain points, and ultimately lost customers who never return.

Traditional manual review reading is slow, subjective, and inconsistent. One manager might notice pricing complaints; another might focus on atmosphere. There's no systematic, reproducible method for turning reviews into strategic insight.

The Solution

AI-powered Natural Language Processing (NLP) sentiment analysis automatically reads and classifies every single review as Positive, Negative, or Neutral — in seconds. Each review is also tagged with a topic category such as Food Quality, Wait Time, Pricing, or Service, so owners can see exactly what is driving satisfaction or frustration.

The result is a complete marketing intelligence dashboard with zero coding required. Instead of guessing, a small business owner can immediately see their top strengths, their most urgent problems, how satisfaction changes month-over-month, and how they stack up against competitors — all from a single automated AI pipeline.

This project demonstrates how accessible AI tools like ChatGPT and Claude can be used by non-technical business owners to make data-driven marketing decisions.

🧠

Sentiment Analysis (NLP)

Natural language models classify each review as Positive, Negative, or Neutral with high accuracy.

🏷️

Topic Extraction

AI automatically tags each review with categories: Food, Service, Pricing, Wait Time, Ambiance, and Menu.

📊

Data Visualization

Interactive charts and dashboards turn raw classified data into clear, actionable marketing intelligence.

How It Works — The AI Pipeline

1

Collect Reviews

Google & Yelp
75 reviews
Oct 2025–Mar 2026

2

AI Sentiment Classification

ChatGPT/Claude NLP
Positive / Negative / Neutral

3

Topic Extraction

Food Quality, Pricing, Wait Time, Service, Ambiance, Menu

4

Visualize & Report

Charts, Dashboard, Marketing Recommendations

Tool / Platform Role in Pipeline Stage
ChatGPT / Claude AINLP sentiment classification and topic extractionAnalysis
Microsoft ExcelData organization, storage, and preliminary tallyingProcessing
React + Recharts / Chart.jsInteractive data visualization and dashboardVisualization
Google ReviewsPrimary source of customer review dataCollection
YelpSecondary source of customer review dataCollection

What the AI Found — Key Results

Sentiment Distribution

Topic Sentiment Breakdown

Monthly Satisfaction Trend

Competitor Comparison

Top Strength

Food Quality scores 76% positive reviews. Customer Service is even stronger at 89% positive — guests consistently praise attentive, knowledgeable staff.

Critical Issue

Wait Time is 100% negative across all reviews — every single mention is a complaint. Pricing is 86% negative, indicating a serious value-for-money perception gap.

Trend Alert

Satisfaction dropped sharply from 100% in October to 0% in December–January — a holiday season crisis that was completely invisible without AI analysis.

Full Review Dataset — All 55 Reviews Analyzed

ID Review Text Stars Date Platform Topic Sentiment

3 Marketing Recommendations

01

Fix Wait Times

Implement a digital waitlist app with SMS updates so guests can wait offsite comfortably. Hire additional front-of-house staff specifically for weekend rushes and the holiday season (November–January). Launch an Early Bird Special — 15% off all orders placed before 6 PM — to redistribute demand and reduce peak-hour congestion. Track monthly wait time review sentiment to measure progress.

02

Reframe Pricing

Launch a prix fixe "Chef's 3-Course for $45" menu that clearly communicates value over individual item pricing. Introduce a loyalty program — every 10th visit earns a complimentary course — to reward repeat customers and increase perceived value. Run a targeted social media campaign promoting the premium local and imported Italian ingredients used in each dish, anchoring the price to provenance and quality.

03

Expand Menu

Add 3–4 new vegetarian and gluten-free dishes to address the menu variety complaints identified in the AI analysis. Clearly label all allergen-friendly items on both the physical and digital menu. Partner with dietary-focused local food bloggers and Instagram creators for organic promotion. This expands the addressable customer base and signals that Bella's Bistro is inclusive and modern without abandoning its Italian identity.

About This Project

BK

Biswas Khatiwada

Iowa State University

Business & Marketing

AI 2010 — Intro to Applied AI

Course
AI 2010 — Introduction to Applied AI
Major
Business & Marketing
University
Iowa State University
Analysis Period
October 2025 – March 2026
Total Reviews
75 (Google + Yelp)
AI Tools Used
ChatGPT, Claude AI, Chart.js
AI does not replace the marketer — it gives them superpowers.

— Biswas Khatiwada, Iowa State University