Webinar: From Narratives to Networks

Transforming Textual Feedback from Consumers into Causal Bayesian Networks with BayesiaLab and Hellixia

Recorded on Thursday, January 30, 2025.

Generative AI is excellent at creating narratives, such as explanations, summaries, and insights in natural language. However, decision-makers require more than just narrative generation; they need mathematical models, including statistical analyses, optimizations, and simulations, to support robust quantitative reasoning and forecasting. This is where Hellixia, BayesiaLab's revolutionary GenAI Partner, comes into play—it effectively connects these two domains.

In this webinar, we’ll demonstrate how Hellixia transforms unstructured textual feedback from consumers, such as responses to open-ended survey questions, into actionable insights. By integrating semantic mapping, probabilistic causal modeling, and collaborative elicitation with multiple LLMs, we uncover the drivers of customer satisfaction.

To showcase these innovative methodologies, we’ll explore two common types of textual sources in automotive market research:

• Part 1: Consumer feedback summaries (e.g., from focus group reports)
• Part 2: Individual auto buyer reviews (e.g., comments from survey forms)

Join us to see how Hellixia improves decision-making by joining complex qualitative narratives with formal mathematical models.

Background & Context

Collecting and correctly interpreting customer feedback is crucial for any company developing new products or improving existing ones. For many Fortune 500 companies, BayesiaLab has become the de facto standard for analyzing consumer surveys. As a strategic partner of Procter & Gamble since 2009, Bayesia has been refining machine-learning techniques to generate Probabilistic Structural Equation Models for Key Driver Analysis from survey data. The characteristics of thousands of products on the market today — as diverse as chewing gum, laundry detergent, or healthcare services — have been refined with BayesiaLab's modeling and optimization tools.

The Bayesia team has been teaching this evolving methodology for nearly two decades by conducting courses around the world, hosting seminars, and writing countless case studies. You can find many examples on this website (see Seminar on Key Driver Analysis, Chapter 8: Probabilistic Structural Equation Models, etc.).

Webinar Agenda

Part 1: Analyzing a Summary of Customer Feedback

In the first part of the webinar, we analyze a textual summary (or narrative), which synthesizes the feedback from owners of the new Solaris 1 — a fictional vehicle brand we use for this study. The summary captures experiences with the car, which are grouped into topics, such as "Exterior," "Interior," "Performance," "Reliability," "Comfort," "Safety," and "Value for Money."

Our objective is to identify the overarching themes and key drivers of overall satisfaction. Using Hierarchical ICI Modeling, we demonstrate the benefits of grouping related dimensions into thematic categories, leading to a structured and interpretable representation of customer opinions.

Example from Part 1: Hierachical ICI Network

Part 2: Analyzing a Collection of Individual Customer Reviews

In the second part, we analyze a different type of text that relates to a different vehicle. This second text is a collection of individual reviews from 50 new owners of the Thunderbolt X — another made-up model. So, each respondent speaks specifically about their personal experience with this car.

Example from Part 2: Semantic Graph with the Extracted Key Dimensions

Common Methodology

Despite the difference in text sources, a synthesis versus a collection of 50 individual reviews, we follow the same methodology to extract dimensions, build a hierarchical causal model, and visualize their impact on overall satisfaction.

Furthermore, we perform a sentiment-driven exploration of user emotions. By isolating emotional dimensions such as High Satisfaction, Excitement, Frustration, we reveal their distinct influence on user perception and overall satisfaction. The combination of Tornado Graphs and Quadrant Charts provides a comprehensive view of user sentiments, highlighting both risks and opportunities.

Whether analyzing aggregated narratives or individual user reviews, the proposed methodologies provide actionable insights to improve product design, marketing, and, ultimately, overall customer satisfaction.

About the Presenters

Dr. Lionel Jouffe

• Dr. Lionel Jouffe is co-founder and CEO of France-based Bayesia S.A.S. Lionel holds a Ph.D. in Computer Science from the University of Rennes and has worked in Artificial Intelligence since the early 1990s. While working as a Professor/Researcher at ESIEA, Lionel started exploring the potential of Bayesian networks.
• After co-founding Bayesia in 2001, he and his team have been working full-time on the development of BayesiaLab. Since then, BayesiaLab has emerged as the leading software package for knowledge discovery, data mining, and knowledge modeling using Bayesian networks. It enjoys broad acceptance in academic communities, business, and industry.

Stefan Conrady

• Stefan Conrady has over 20 years of experience in decision analysis, analytics, market research, and product strategy with Mercedes-Benz, BMW Group, Rolls-Royce Motor Cars, and Nissan, which included assignments in North America, Europe, and Asia.
• Today, in his role as Managing Partner of Bayesia USA and Bayesia Singapore, he is recognized as a thought leader in applying Bayesian networks for research, analytics, and reasoning.
• Recently, Stefan and his colleague Dr. Lionel Jouffe co-authored Bayesian Networks & BayesiaLab — A Practical Introduction for Researchers, which is now available as an e-book.