Jun 03, 2025
How Class Editori used LLM user analytics to build a GenAI assistant users actually want
Class Editori logo over financial newspaper - case study on LLM user analytics with Nebuly

Summary

Class Editori S.p.A. is an Italian media conglomerate based in Milan and listed on the Italian Stock Exchange.

Sector

Media/Financial News

Use case

MF GPT, a newspaper GenAI chatbot

Launch

June 2025

Jun 03, 2025
How Class Editori used LLM user analytics to build a GenAI assistant users actually want

Class Editori’s vision was clear: to extend its trusted financial expertise through a GenAI assistant that could inform, educate, and support users’ financial decisions.

Yet until they arrived at MF GPT, they faced a fundamental obstacle: they didn’t know which use cases would resonate most with subscribers, or how users would truly interact with an AI-powered experience.

We didn’t know what the main use case would be, and that was one of the main challenges we faced. But when we saw the demo, it was clear Nebuly was a complete tool with a lot of capabilities to help us in this process.


AndreaPazzaglia

Andrea PazzagliaHead of AI Development at Class Editori

By integrating Nebuly’s LLM user analytics, Class Editori uncovered that real users prioritized queries on stock-specific details, market movements, and financial instruments. These GenAI insights enabled the team to refine the assistant’s responses and concentrate development on features subscribers truly valued.

MF GPT is the result of that journey: Italy’s first autonomous generative AI system applied to economic and financial news, designed to transform how users access and engage with Class Editori’s in-depth news research and editorial analysis.

With Nebuly’s analytics platform providing a clear view into user behavior, Class Editori achieved measurable improvements in user adoption and engagement, leading to these key outcomes:

50%
faster response
times
Clear
use case
definition
Actionable
insights
on user interests

The challenge

Building a GenAI assistant to deliver trusted financial insights required a clear understanding of how real users would interact with the system. Class Editori faced three key challenges:

01. Unclear use cases: identifying which features subscribers would value most in a GenAI assistant.
02. System correctness and reliability: reducing errors and improving stability to ensure trustworthy responses.
03. User engagement and retention: understanding how to capture and sustain user attention to drive adoption.

Without insight into real user behavior, Class Editori risked launching an assistant that wouldn’t meet subscribers’ needs or achieve their business goals. That’s why tracking GenAI user analytics became a mission-critical priority.

The good part is that we’re building the tool with Nebuly - and that’s not something to take for granted. If we had started the implementation without it, we would have missed many important aspects.


AndreaPazzaglia

Andrea PazzagliaHead of AI Development at Class Editori

The Nebuly approach

To tackle these challenges, Class Editori partnered with Nebuly from the outset, building their GenAI assistant with integrated LLM user analytics to capture real user behavior in real time.

By tracking exactly what subscribers were searching for, the team could prioritize features that aligned with real user demand - avoiding investment in unnecessary capabilities.

Nebuly’s dashboards provided instant visibility into error patterns, enabling the team to quickly identify and resolve issues, boosting the AI’s stability and reliability.

We are building and developing while we observe the data in near real time, making a lot of adjustments as we go.


AndreaPazzaglia

Andrea PazzagliaHead of AI Development at Class Editori

Building the Future of GenAI Engagement

Class Editori also leveraged Nebuly’s analytics to lay the groundwork for smarter content and product decisions. They plan to share usage insights with journalists to better align editorial topics with reader interests, and to run A/B tests on different AI configurations to continuously optimize performance and engagement.

In the future, the plan is to use the tool to make A/B tests on different configurations.


AndreaPazzaglia

Andrea PazzagliaHead of AI Development at Class Editori

With Nebuly’s support, Class Editori transformed their GenAI assistant from a promising idea into a data-driven solution optimized for user engagement, delivering faster, more relevant, and more reliable experiences for their subscribers.

Ready to unlock LLM analytics for your business?
Join leading enterprises using Nebuly to turn conversations into intelligent insights.
Our LLM analytics platform empowers your team to understand, optimize, and protect AI-driven user experiences.