Skip to content
Back to overview

AI Training / Enablement / Workshop

STIHL – AI Enablement with Use Cases from Shopfloor to Warranty

AI enablement for STIHL: more than 30 employees from production, engineering, logistics and office developed their own AI use cases, from a RAG knowledge assistant to warranty data analysis.

STIHL – AI Enablement with Use Cases from Shopfloor to Warranty

Starting point

More than 30 participants from production, engineering, logistics and office, many of them with hardly any prior exposure to AI. Across all areas, the same three problems surfaced:

  • Downtime and rework cost money, for example when knowledge depends on individual people.
  • Knowledge is not systematically available. Whoever wants to know something asks colleagues.
  • Duplicate work due to tool breaks between different systems and paper. Analyses such as warranty data take correspondingly long.

Solution

Two compact sessions in one day, in cooperation with the IHK (Chamber of Industry and Commerce). The focus was on working with real processes:

  • AI fundamentals, strengths, limits and hallucinations
  • Two purpose-built demo tools: a CustomGPT based on operating manuals and image generation with Flux 2 Pro for product-related visuals
  • Automation demos with n8n, Power Automate and Lovable
  • 30 minutes of group work on the participants’ own pain points

The groups came up with three concrete implementation ideas:

  1. Shopfloor knowledge assistant: setup instructions and fault diagnostics on a tablet terminal, technically a RAG system on the company’s own documentation
  2. Warranty intelligence: pattern recognition in warranty returns with forecasts for spare-parts planning
  3. Personalized search agent for the intranet, a typical use case for AI agents (German)

Technology stack

  • LLM tools in practice: ChatGPT, Claude & co. (prompting, limits, hallucinations)
  • CustomGPT based on operating manuals
  • Image generation with Flux 2 Pro
  • Automation: n8n, Power Automate, Lovable
  • RAG concepts for knowledge assistants on the shop floor

Result & Impact

  • More than 30 employees from four areas in two sessions on one day
  • Two tailored tools instead of generic demos
  • Three concrete use cases from the business units. None of them had been on an IT list before.
  • Clear levers identified: warranty data analysis that currently takes weeks, and machines staffed by 2 people instead of 1