Rosiglioni Impianti — Product Identity & Operational Infrastructure

Our entry didn't start with a brief. It started with a harder question: before communicating better externally, can you speak clearly internally?

The Decode phase involved quantitative and qualitative analysis of the existing catalog — structure, classification logic, data completeness, description consistency, distribution by category. The goal wasn't to find defects: it was to understand where the problems were and what their nature was.

What we found:

• Non-uniform codes between categories and between suppliers — each part of the catalog had developed its own conventions
• The SPEC field filled with GEN (generic) in 71% of cases — data that carries no useful information
• Prices and stock not present in the catalog, managed separately in the management system
• No explicit management of obsolete or replaced products — they lived in the catalog as active items
• An "Accessories" category with 2,763 items — signal that the system stopped classifying and started accumulating
• Classification knowledge existed only in people, not documented anywhere

The analysis produced a precise map of where the problems were and what their nature was. Some categories were well maintained; others nearly abandoned. Some suppliers had dozens of unconsolidated variants. The picture was clear.

The new architecture couldn't be designed from outside alone. Every structural decision required dialogue with the team — because knowledge of the technical domain of medical gases is specialized, and classification choices have real operational consequences.

Together with Sabrina Rosiglioni, COO of the company, we redesigned the catalog structure from within through structured work sessions. Every choice was stress-tested: does this category hold when the configurator needs to automatically select components? Is this code memorable for someone who's never seen it? Does this distinction make operational sense, or is it just historical?

• New MACRO taxonomy in 12 categories based on the actual technical function of products in the system — not supplier origin
• Shared and documented supplier abbreviation system — 24 main suppliers each assigned consistent 3-letter abbreviations
• Explicit management of obsolete items with -OLD suffix — visible in the system but identified, not lost
• Dedicated category for patient-bedside systems, built on language already in use by the team
• PENDING sheet for items still to be classified — transparency on work status

Non-obvious decisions were documented explicitly — with the reasoning and rejected alternatives. A knowledge asset worth as much as the catalog itself: whoever enters the company two years from now will understand why the system is built that way.

The new SKU coding logic was designed around one precise principle: an item must be identifiable by someone who has never seen the old system. Each component of the code carries information — not just distinction.

The final system combines: functional category, supplier abbreviation, sequential number. The result is a code readable aloud, shareable by phone or email, memorable for sales staff. Not glamorous. Functional.

In parallel, the internal operational guide was built: a document that explains the classification logic, coding rules, obsolete management and the process for adding new items. Designed for someone who joins the company two years from now and has never seen the old system.

• Coding logic with three-component structure (category + supplier + sequential)
• Consistent naming rules for product families
• Variant management with standardized suffixes
• Operational guide for personnel onboarding and ongoing catalog maintenance

A reorganized catalog without economic data is a half-tool. The Deploy phase involved integrating real management system data into the new structured catalog.

Each item in the new structure was enriched with: updated purchase price, current stock, unit of measure, relevant operational notes. The source is the company management system — data was not reworked, it was transferred into the correct structure.

In this phase, AI was also used in a targeted way to accelerate mechanical work while maintaining human judgment where necessary:

• Inference of the correct SPEC from text descriptions for 104 items classified generically
• Identification of substitution patterns in legacy codes and internal notes
• Extraction of commercial category (component, spare part, kit, consumable) from descriptions where readable

Classification decisions, validation of economic data and handling of ambiguous cases remained human. In a field where a wrong item in a medical gas installation has real consequences, precision is non-negotiable.

The project was born with a concrete objective in sight: Rosiglioni Impianti wants to build a digital configurator for designing medical gas equipment systems. A tool that allows sales staff — and in future customers — to select the correct components, generate coherent specifications, produce structured quotes.

That configurator couldn't have relied on the previous catalog. Every automatic selection would have been unreliable. Every generated quote would have required manual verification. The diagnosis and reorganization of the catalog were the prerequisite — not the accessory.

With the new data foundation ready, Rosiglioni now has the groundwork to build digital tools that actually work. The governed catalog is the zero point of the company's digital evolution.

Ongoing support continues: progressive enrichment of descriptions from PDF technical sheets from suppliers, expansion of data coverage, evolution of the operational guide with the team.