If you’re trying to move AI from scattered experiments to repeatable adoption, REHAU’s experience shows a practical sequence: make experimentation safe, build the support system, then scale usage.

Nils Wagner, CEO of REHAU New Ventures, shared how this has worked in practice during our recent EUVC Corporate AMA.

After a January soft launch, REHAU’s internal AI platform reached 500 users by April and 1,182 by August 27, significantly more than 10% of its installed user base.

According to Nils, adoption had been driven almost entirely by employee pull. REHAU only began promoting the platform more actively once the support infrastructure was ready.

Here are three parts of its operating model that corporate innovation and transformation teams can apply.

After seeing weak adoption of Microsoft Copilot, REHAU introduced a browser-based AI environment that passed its IT review, met its enterprise security requirements and gave employees access to multiple large language models.

Put it into practice: Give employees a clear route from identifying an opportunity to testing it in an approved environment. Resolve access, security and compliance friction before making adoption a target.

REHAU supported adoption through:

It also built an internal AI Academy in about eight weeks, producing approximately 50 hours of interactive content covering everything from basic use to workflow development.

The resulting pull was tangible. Internal information lunches run by divisional AI leads regularly generated another 20 to 30 licence requests.

Put it into practice: Before promoting an AI tool, ask whether you could onboard 100 new users, answer their questions and help the strongest use cases progress.

More complex, sensitive or costly workflows are routed to divisional use-case boards. These assess risk, regulatory requirements, expected costs and estimated savings, while workflows involving core systems require additional IT involvement.

Nils described one employee spending approximately 10 hours building an agent for complex tax-filing requirements. He estimated that comparable work from an external consultant would have cost around €100,000.

That is an early example rather than proof of programme-wide ROI. Nils said REHAU still needs more operating data to quantify overall productivity gains and cost savings.

Put it into practice: Create a route for promising experiments to become governed use cases, with clear ownership, risk checks and an expected-value calculation before additional resources are committed.

Nils frames AI implementation across three horizons:

REHAU’s experience suggests that scaling AI adoption is not simply about putting more licences into employees’ hands. Safe access, enablement and governance need to be ready before wider promotion creates more demand.

EUVC Academy brings these kinds of practical operating discussions to small-group AMAs and masterclasses with investors, operators and corporate innovation leaders doing the work, with members receiving priority access to upcoming sessions.