AI potential everywhere, and no room for loose ends
Waldemar Link GmbH & Co. KG (LINK) is a family-owned German manufacturer of orthopaedic implants. It operates in one of the most tightly regulated industries there is. Every system that touches data has to stand up to scrutiny from IT, compliance and regulators.
LINK saw strong potential for AI across regulatory, R&D, marketing, sales, finance and master data. The question was never whether AI could help. It was whether LINK could roll it out broadly without creating a governance problem it would later have to unwind.
Challenge
Before putting AI in front of the whole organisation, LINK set three non-negotiable conditions.
- Data had to stay within the EU. As a regulated medical device manufacturer, LINK could not accept inference running outside the EU.
- Clear control over who uses what, and at what cost. Broad availability only works if access, model choice and spend can be managed per team. Nobody should discover them after the fact on an invoice.
- Full transparency for IT and compliance. Every use of AI needed to be visible and accountable, not scattered across unmanaged tools and personal accounts.
The ask was a single, governed platform that would make Claude available to everyone, and a foundation departments could build on safely.
Solution
Firemind supported LINK from strategy through to operations, in four connected pieces of work.
Executive alignment on where AI creates value
The CFO and department heads took part in an executive workshop to identify and prioritise the use cases that mattered most. The rollout started from business priorities, not technology.
A secure AI gateway as the single entry point to Claude
Every request to Claude runs through one gateway. That gives LINK central control over access, team budgets, guardrails and model choice, with EU inference throughout.
Claude where people already work
Firemind rolled out Claude Cowork and Claude in Microsoft 365, integrated with LINK’s existing identity and device management. Employees sign in with their existing corporate identities and use Claude directly in Outlook, Word, Excel and PowerPoint.
A governance framework that lets departments build for themselves
A governance framework and AI development guidelines give departments clear rules for building new use cases safely. Ideas no longer have to go through a central team.
What the platform gives LINK now
Beyond the figures above:
- Claude in the daily work of every department. Employees use it in Cowork and directly in Outlook, Word, Excel and PowerPoint.
- Every interaction governed centrally. All usage runs through EU inference, with cost transparency per team and access based on existing corporate identities.
- Departments building, not just using. The first departments were building their own use cases on the platform after just two weeks.
- A foundation ready to scale. The platform will serve business users and engineers, and it will also be the foundation for the agentic applications LINK builds next.
For us, the key was to make AI broadly available across the organisation without compromising on governance, security or cost transparency. Together with Firemind, we moved from concept to a scalable platform within just a few weeks and are connecting it to our internal systems and data sources, enabling our teams to develop and use AI in a controlled environment. This gives us a strong foundation to expand AI across the LINK Group while maintaining the level of control we require as a regulated medical device company.
Scope: this reflects the platform rollout from kickoff to production and adoption to date. The next stages of the programme are connecting the platform to LINK’s internal systems and data sources, and building agentic applications on it.
