This website uses cookies so that we can provide you with the best user experience possible. Cookie information is stored in your browser and performs functions such as recognising you when you return to our website and helping our team to understand which sections of the website you find most interesting and useful.

Client
One of the European manufacturing sites of a major pharmaceutical company, specializing in the industrial production of solid dosage forms (e.g., tablets and capsules), was seeking a way to reduce response times during emergency situations. Due to the large scale of operations and complex machine infrastructure, the maintenance team needed a tool that would enable fast and secure access to knowledge—regardless of shift, tenure, or employee experience level.
Challenge
A machine failure in a manufacturing facility demands an immediate response. Unfortunately, access to critical information was fragmented:
- Technical documentation existed in multiple formats,
- Failure history was stored across separate systems,
- Operational know-how was undocumented and “scattered across shifts.”
As a result:
- Response times were too long,
- Diagnoses were often based on intuition or incomplete data,
- The service team spent valuable time searching for information or consulting colleagues,
- Potential solutions were frequently repeated or reinvented from scratch.
The goal of the project was to create a system that:
- Integrates documentation, incident history, and technical notes in one place,
- Allows employees to ask questions in natural language and receive accurate answers,
- Is available 24/7 and easy to use, even for non-technical staff,
- Complies with IT security policies and can be deployed on-premises.

Solution
In response to the client’s needs, we developed Herodotus – an AI-powered Help Desk system designed to support maintenance operations and simplify the daily work of technical teams.
Herodotus is built on a Retrieval-Augmented Generation (RAG) architecture. Unlike traditional AI models that generate answers from scratch, Herodotus doesn’t “make things up.” Instead, it searches through documents, incident tickets, and other approved resources, then formulates answers to user questions—clearly referencing the specific sources it’s based on, such as PDFs, tickets, or system logs.
Key Features of the System
- Manufacturer Documentation Search
Users don’t need to know file structures or folder names—just ask a question in natural language. The system locates the exact section of a manual or technical schematic. - Ticket Database
All past incidents and their resolutions are stored in one place, ready to be reused whenever a similar issue arises. - Integration of Multiple Sources
Herodotus connects documentation, repository files, service bulletins, checklists, and logs—eliminating the need for manual searches across systems. - Barrier-Free Shift Work
The system is available at all times, across all shifts—ensuring every user accesses the same, up-to-date knowledge base. - Secure Deployment and Quality Control
Herodotus operates within a “restricted context”—it processes only client-provided data and never generates answers from outside the defined knowledge base.

Results
Faster Diagnosis and Problem Resolution
Employees can identify the root cause of a malfunction more quickly—even without prior experience with the specific machine.
Retention and Accessibility of Operational Knowledge
Knowledge doesn’t disappear when employees leave—it’s archived, searchable, and accessible to the entire team.
Reduced Downtime and Increased Machine Availability
Faster response times and accurate diagnoses directly translate into shorter equipment downtime.
Secure AI Deployment in a Production Environment
The system relies solely on internal documentation and adheres to the company’s IT security policies. It can be deployed either on-premise or in the cloud (e.g., Azure).
Collaboration Model
Full Deployment Tailored to the Plant’s Reality
We begin with an in-depth analysis of available resources—technical documentation, incident databases, work files, and procedures. Based on this, we configure the system to fit the specific processes and needs of the maintenance team.
Team Training
We provide hands-on training for technical staff. We teach not only how to ask questions to the system, but how to apply it effectively during shifts—for diagnosing faults, solving issues, and working with documentation.
Ongoing Support and Development
As part of our support package, we offer user onboarding, prompt optimization, and advisory services for further solution development. We assist in expanding the knowledge base, scaling the system to additional departments, and integrating it with the existing infrastructure.
About the Project
Herodotus doesn’t replace people—it empowers them. It gives teams access to knowledge that was once buried in files, notes, and memory. It enhances decision-making, speeds up response times, and organizes operational know-how—right where every minute counts.

Shape Recognition Automation Using Machine Learning – Optimizing Production Costs at a Toronto Plant
Design, Development, DevOps or Cloud – which team do you need to speed up work on your projects?
Chat with your consultation partners to see if we are a good match.
