24th FVI Expert Breakfast

Topic: "Knowledge to go" – How Enercity digitizes expert knowledge before retirement comes.

In this session, Markus and Björn (Enercity) together with Sina showed how to pack 40 years of experience into an app.

• The trigger: The heat transition forces Enercity to double the district heating network. At the same time, the team doubles, but the experienced experts are retiring. The problem: How do I teach a 25-year-old electrician in 1 year what the 60-year-old learned in 30 years?
• Documents are not enough: Sina warned against simply throwing all PDFs into an AI ("Dump & Pray"). AI can quote from manuals, but it cannot "think logically" with complex errors. The valuable knowledge is not in documents, but in the heads ("If it rattles there, it's usually valve X").
• The "heatmap" of knowledge: Enercity uses an app that guides the technician: "What do you see? What do you hear?" (Symptoms). Behind it lies a matrix (heatmap) that links symptoms with causes. This matrix is fed by the experts. This is not AI magic, but structured experiential knowledge.
• Human-in-the-Loop: An important point in the discussion: Can the AI learn unfiltered? No. Enercity relies on "curators" (experienced employees) who check the technicians' feedback before it flows into the knowledge base. Otherwise, the system learns nonsense ("Garbage in, Garbage out").
• ROI through safety: Markus Felden emphasized that the ROI is hard to measure in euros, but in safety. If the technician knows thanks to the app: "Caution, there is still pressure here", he comes home healthy in the evening. That is priceless.

Classification: ADAM as a knowledge vault

This episode is the perfect confirmation of our vision of the "Digital Vault".

• ADAM goes further: We connect the diagnosis with the asset history (When was the last maintenance?) and the spare parts (Is the valve in stock?). We offer the context that a pure diagnosis app often lacks.
• Democratizing knowledge: Björn (Enercity) reported that the young colleagues are grateful for the app because it provides security. ADAM does exactly that: It is the "digital mentor" that is always there. This makes the job attractive for newcomers (see Ep 14).
• The curated approach: We share the view that AI should not learn "wildly". In ADAM, there are role concepts: The expert releases knowledge, the trainee consumes it. This ensures quality.

Conclusion: Knowledge is the new gold. Those who do not mine (extract) and refine (curate) it lose their most valuable treasure to retirement.