Charles AI
Intelligent Insight Engine for Automotive Value Data
Context & Objectives
As an extension of the Value App platform, we developed and integrated Charles, an AI-powered engine designed to extract meaningful insights from structured vehicle assessment data. The goal was to support product, UX, and quality teams in identifying patterns, anomalies, and actionable feedback across large volumes of evaluation data
Gemini 2.5 Integration with Value App
Charles was built directly into the existing Value App ecosystem, leveraging the rich, structured database of vehicle assessments generated by global teams. Powered by Gemini 2.5, the system was designed to operate within our client’s internal environments, respecting data governance and security requirements.
Capabilities & Features
Charles uses a combination of machine learning models and rule-based logic to surface key insights from thousands of assessment points. It helps teams detect inconsistencies, highlight standout vehicle attributes, and accelerate comparative analysis across models and project stages.
Human-Centered AI
We focused on making AI outputs transparent and actionable for teams that may not have a technical background. The interface was designed to clearly communicate insights with contextual explanations, and to integrate smoothly into the team’s existing workflows within the Value App.
Outcomes
Charles enabled faster, more informed decision-making by automating a large part of the analysis process and offering real-time guidance during assessments and reviews. It serves as an intelligent assistant embedded in the product development pipeline.
