Carglass enhances customer sentiment analysis with generative artificial intelligence and Tableau

The project

Carglass is the leading windscreen and auto glass repair and replacement company. With an extensive network throughout Italy and a strong focus on customer experience, Carglass constantly invests in improving the quality of its services, with thousands of online reviews and feedback from post-service surveys. The company wanted to transform the textual data into structured insights to better understand customer sentiment and quickly find areas in need of improvement, locally and nationally.

Carglass asked Atlantic Technologies to support its analytics team with analysing customer reviews on Trustpilot and Google in a project using Tableau and OpenAI.

The dataset, consisting of ratings and text of reviews, was processed through the Chat GPT API for:

• Attributing a sentiment to each feedback as positive, negative, or neutral;
• Extracting keywords and recurring concepts;
• Generating automatic summaries of reviews;
• Classifying each review into thematic clusters defined earlier by the company.

The data were integrated into Tableau interactive dashboards, in order to analyse performance by location, and comparing geographic areas in real time.

The benefits

The project showed the analytics team a new way to read and analyse data. Carglass can now rely on:

• Strategic insights from textual data along with a greater understanding of the real perceptions of clients;
• AI applied to Business Intelligence to conduct analyses, classifications, and more accurate synthesis on feedback;
• Tableau dashboards that are always up-to-date, intuitive, and navigable at multiple levels;
• Greater overview by geographic area and operational centres.