How Oude Reimer made every technician as knowledgeable as its most experienced
A 10-person service team relied on memory and manual PDF searches to troubleshoot machines from 33 manufacturers. Lleverage built an AI knowledge base that indexes every manual automatically and returns structured answers in 70 seconds.
When expertise lives in people's heads, it doesn't scale
In technical service, the knowledge base never stops growing. Every new machine model adds another set of manuals. Every completed repair adds another set of lessons. But the way technicians access that knowledge rarely keeps pace. It stays locked in documents nobody reads cover to cover, in the heads of a few experienced people, and in service reports filed away after the job is done.
Oude Reimer distributes and services precision machinery for the steel industry across the Netherlands. Their 10-person service team — seven field technicians and three inside service staff — handles between two and ten customer inquiries per day, covering machines from 33 different manufacturers. Each inquiry requires pulling from a combination of OEM manuals, written in German or English, and historical service reports known internally as Werkbonnen.
The manuals and service reports were scattered across SharePoint with no structured search. A technician looking for a specific troubleshooting procedure had to know which manufacturer folder to check, which manual to open, and roughly where in the document to look. For unfamiliar machines, that meant scrolling through hundreds of pages of technical German — or calling a colleague who might know the answer.
"Nobody can memorize every manual from 33 different manufacturers. When a customer calls with a machine problem, the quality of the answer depends on who picks up the phone and what they happen to remember."
— Hugo Oude Reimer, Director, Oude Reimer
There was no after-hours self-service, no standardized way to access technical knowledge, and no mechanism to transfer what experienced technicians knew to newer team members. The knowledge existed. It just was not accessible in a way that scaled.
From SharePoint folders to a searchable knowledge base
Milos Mandic, Forward Deployed Engineer, and Badr Eddial, CTO at Lleverage, worked with Remco Hooft, Oude Reimer’s Technical Owner, to build a two-part system: an automatically maintained knowledge base and a chat assistant that sits on top of it.
The knowledge base draws from two sources. OEM manuals are ingested directly from SharePoint — when a new manual is uploaded to the right folder, the system automatically chunks it, indexes by manufacturer and machine model, and adds it to the searchable library. Historical service reports follow the same path. The system currently holds 170 manuals from 33 manufacturers alongside 144 historical service reports.
The auto-ingestion pipeline was a deliberate design choice. A knowledge base that requires manual uploads goes stale the moment someone forgets to add a new document. By connecting directly to SharePoint, the system indexes every new manual or service report automatically — zero manual effort, minutes of lag. The knowledge base grows as the business grows, without anyone maintaining it.
The chat assistant gives technicians a structured way to query this library. A technician selects a manufacturer and machine model through an intake form, then asks a troubleshooting question in plain language. The system retrieves relevant sections from the matching manuals and service reports and returns a structured answer with specific references to the source document and section.
Every technician answers like the most experienced one
The chat assistant went into production in December 2025. All 10 technicians on the service team now use it as part of their troubleshooting workflow.
The results are consistent. Zero errors across all sessions since launch. Average session duration: 70 seconds — the time from asking a question to receiving a structured, referenced answer. What previously required manual searches across hundreds of pages of technical documentation now takes just over a minute.
"The answers come back with references to the exact manual and section. That is what builds trust — the technician can verify the information themselves before passing it to a customer."
— Remco Hooft, Technical Owner, Oude Reimer
What made this work was not the chat interface itself, but the knowledge base underneath it. The auto-ingesting knowledge base means the system improves every time Oude Reimer adds a new manual or completes a service report — without anyone actively maintaining it. The chat assistant is the interface. The continuously growing, self-maintaining knowledge base is the asset.
Looking ahead
The chat assistant is the foundation for a broader automation roadmap at Oude Reimer.
Quote rebranding automation. When Oude Reimer receives a supplier quote, it currently requires manual reformatting to match their own branding and layout standards — a process that can take up to half a day per quote. A new workflow is in development to automate roughly 80% of that reformatting, reducing per-quote effort to five minutes.
Price list automation. Oude Reimer currently dedicates two days per week to manually transforming OEM supplier price lists into Business Central import files — updating prices, adding new articles, and blocking discontinued ones across thousands of SKUs. A new workflow is in development to automate this end-to-end, including German-to-Dutch product description translation.
Customer-facing self-service. The current system serves the internal service team. A future phase will expose a version of the chat assistant to customers for common technical inquiries — extending 24/7 knowledge access beyond the service desk.