Introduction
Zerynth is an Italian deep-tech company that develops an IIoT platform for the rapid and scalable digitalization of industrial production. The company is currently evolving its solution into an Industrial AI Copilot platform, equipped with a conversational agent that enables natural interaction with production data and processes. The development of this artificial intelligence component is considered strategic for the evolution of the product and the company’s future positioning.
The Problem
Modern factories are gold mines of data.
- Modern factories are goldmines of data. IoT sensors capture every vibration, energy consumption figure and production cycle. However, this information often remains ‘silent’ or accessible only to those who know how to interpret complex software.
- Added to this is the problem of documentation: user manuals, standard operating procedures (SOPs) and technical guides are often confined to static PDF files running to hundreds of pages, which are difficult to consult when needed.
The Solution: A Copilot based on Agentic AI and RAG.
At the heart of the project is an advanced digital assistant that does not merely ‘respond’, but ‘reason’ about the data. Unlike traditional chatbots, this system uses an Agentic RAG (Retrieval-Augmented Generation) architecture.
What does it mean in practical terms?
- Real-time data access: The system queries the platform’s APIs directly to retrieve KPIs on machinery, consumption and efficiency (OEE).
- Intelligent document search: Using a vector database, the Copilot ‘reads’ PDF manuals and provides precise answers, citing the source and the exact page.
- Memory and Context: The agent remembers the conversation. If you ask, ‘Compare it with yesterday’, they know exactly which machine or value you are referring to.
Tecnologies
The experimentation has focused heavily on data sovereignty. The core engine is based on Azure OpenAI models to ensure compliance with European regionalization (EU data privacy), orchestrated through advanced frameworks such as Agno and LangGraph.
One of its key strengths is XAI (Explainable AI): each response generated is not a ‘black box’, but includes transparent references to the content used to formulate it, ensuring that the technician can always verify the accuracy of the information provided.
Results from the field: from KPIs to maintenance
During testing, the platform’s integration made it possible to cover critical use cases:
- KPI analysis: Immediate answers on line availability, performance, and quality.
- Technical Support: Automatically opening support tickets via chat when the system detects anomalies.
- Energy Efficiency: Comparative analysis of consumption between different periods to identify waste.
The figures speak for themselves: the ability to reduce machine downtime (by up to 70 per cent in optimised environments) and to drastically cut the time taken to find technical information makes AI an indispensable tool for the competitiveness of manufacturing SMEs.
Benefits for the Company
The adoption of an Industrial AI Copilot is not just a technological upgrade, but a paradigm shift:
- Data democratisation: Information is available to everyone, not just IT experts.
- Decision-making speed: Decisions based on firm data, obtained in seconds.
- Reducing human error: Step-by-step guides and quick access to instruction manuals.
The AI-MATTERS project demonstrates that artificial intelligence in the factory does not have to be complicated. In collaboration with Zerynth, we have created a bridge between man and machine, where natural language acts as the remote control for production.
The smart factory is the one that can answer the questions of its operators. And today, it has finally begun to do so.