Fashion company Laagam open-sources AI system for corporate knowledge management

What was the return on the last pop-up? Which deals were finalised in the meeting with a supplier three months ago? What observations did the sales team relay from their latest conversation with a key client? In most companies, answering such questions still requires a fragmented search. Information is typically scattered across documents, emails, management tools and conversations between departments.

This everyday problem was precisely what led Diego Arroyo, chief executive officer of the Barcelona-based firm Laagam, to develop an infrastructure capable of turning corporate knowledge into a real-time accessible resource. The system centralises information from across the entire organisation. It allows users to query the history of decisions and operations using natural language. It also links data from different areas and executes actions immediately, from generating reports to drafting contextualised commercial communications.

Now, the company has decided to release this technology as open source. This move transcends the fashion industry, a sector historically driven by creativity and product. Increasingly, competitive advantage no longer depends solely on having information, but on the ability to transform it into actionable knowledge.

95 tools, 3,000 documents, zero engineers

According to the executive's own LinkedIn account, the company has been operating for months on what he describes as an “AI operating system”. This architecture connects 95 digital tools, processes more than 3,000 internal documents and executes operational tasks automatically by department. The main engine is Claude, Anthropic's artificial intelligence model. The team has fed it with years of their own corporate memory, all without a single engineer on staff.

Cases like this are currently proliferating at a speed that would have seemed exaggerated just a year ago. Since large language models became accessible without needing to write code, a team without a highly technical background can connect its internal data to an AI model and achieve real operational results. This is possible simply by having a clear idea of the problem to be solved. What once required a technology department now primarily requires judgement.

Fashion companies that have accumulated data for years—including emails; meeting transcripts; sales histories; and supplier communications—are beginning to understand that this disorganised archive is actually their most valuable asset. The key is not the AI model, which anyone can use, but the proprietary context that feeds it. That context is non-transferable and very difficult to replicate.

631 euros per month

However, Arroyo himself is honest about the real scope of the impact. The system has not directly boosted sales. Its value lies in freeing up decision-making time for the management team by absorbing the routine operational workload. It does this for 631 euros per month, the cost of maintaining the AI infrastructure.

Laagam has published the complete system openly under the name Compai. It includes the 53-chapter manual, templates by department and the technical architecture, so that other SMEs can replicate it.

This article was translated to English using an AI tool.

OR CONTINUE WITH

Artificial Intelligence

Laagam