Installers need to work with growing system complexity
Solvis builds heating systems for domestic housing, with deep roots in traditional heating and solar technology and a growing focus on heat pumps. As products evolved, so did the complexity installers faced in the field.
Technical knowledge already existed. Manuals were detailed and accurate. But installers often needed answers quickly while working, not after reading through dozens of pages.
We have manuals up to about 90 pages. You don’t really go through all the manuals while installing a system. Every assistance that helps installers get the right information faster is a help.
The result was a familiar operational challenge:
- installers spending time searching instead of installing
- support teams answering recurring technical questions
- expertise concentrated in a few individuals
- increasing pressure during seasonal peaks
Solvis started looking for a way to operationalize its existing knowledge so installers could access verified answers in real time.
Starting with the problem, not the technology
The initial idea was to build something for end customers only. But early collaboration quickly revealed where even more impact would be.
By mapping available documentation and daily support requests, the project shifted toward installers as the primary users. The reasoning was simple: installers manage the most complex information and face the highest operational pressure. Instead of launching a standalone tool, the goal became clear: structure existing service knowledge and integrate it into the workflows installers already use.
We realized the impact would be greater if we focused on installers.
Turning documentation into operational infrastructure
The first phase focused on organizing knowledge rather than building interfaces.
Solvis and Chapter worked together to:
- structure technical manuals across product lines
- define clear data organization standards
- improve how information was labeled and retrieved
- ensure answers linked back to their source material
One early learning was decisive. Simply pooling documents together did not work. As data became more structured, answer quality improved. Installers could ask technical questions and receive responses grounded in source documentation, supported by snippets and direct references so they could verify answers immediately.
Building trust through verification
For technical environments, reliability matters more than novelty.
A key capability was showing where answers came from. Instead of presenting information without context, installers could see supporting snippets and validate the source themselves. People at Solvis mention that these these information snippets really help. It gives you context it wouldn't help us. You can easily see where the answer comes from which you to much needed context.
This transparency changed how teams viewed AI. It became a structured guide through documentation rather than a black box.
Outcomes and operational direction
The system continues to evolve, and early signs already point toward measurable operational value:
- installers find technical data faster
- support teams can focus on complex cases
- knowledge becomes easier to scale across teams
- recurring users show growing adoption
I think up to 50 percent is realistic as the systems improve. You will always need experts. But AI helps us scale, especially when work peaks.
Since launch, the assistant has handled thousands of technical questions from installers, most of them documentation lookups that would otherwise require scanning manuals of up to 90 pages. This has already redirected hundreds of hours back to fieldwork. Looking ahead, Solvis expects AI to handle an even more substantial share of incoming support questions.
A partnership built on iteration
The collaboration worked because it stayed practical. Weekly meetings, structured project management, and continuous adjustments allowed the system to evolve with real feedback. One thing that really struck Claas was that Chapter had a clear agenda from the start. There were weekly progress meetings. That really helped everyone to get things done. The relationship was less about deployment and more about shared problem-solving:
- refining data structure
- adjusting workflow placement
- learning from real installer behavior
- preparing future integrations
By embedding verified knowledge directly into their workflows, installers work with more confidence, support teams focus on complex cases, and expertise scales beyond individual specialists. The next phase deepens integration. Connecting the system to ERP data will support spare part identification and enable more agentic workflows in the field.
Chapter is also being introduced during Solvis' installer roadshow across Germany to drive engagement and gather field feedback. Recent product breakthroughs have significantly reduced implementation time, making it easier to scale across more data and use cases. One example is a dedicated training agent connected to video and training content, helping installers access learning materials while they work and embedding knowledge directly into the job.
Friendly people, good experience, good project management, professional and practical.
