Switzerland tests a Microsoft 365 exit on 3,000 seats

On September 2 the Swiss Federal Chancellery announced a program to put roughly 3,000 federal employees on a sovereign open-source workplace by the end of 2027, at an estimated CHF 9 million for the first phase. It follows a proof-of-concept called PoC BOSS that put 172 staff on the stack and wrote down what worked.

That second part is the useful artifact. Most Microsoft 365 exit stories are press releases. This one is a test report.

What actually happened

PoC BOSS was built on openDesk, the workplace suite owned by ZenDiS, Germany's centre for digital sovereignty. Nine open-source components, all browser-delivered: Collabora Online for documents, Open-Xchange for mail, calendar, contacts and tasks, Nextcloud for files, Nordeck/Jitsi for conferencing, Element for chat, Univention for identity and access.

The verdict: suitable for key standard processes. Document editing, file storage, collaboration, mail and calendar all rated well with the 172 testers. The Chancellery also framed it as a fallback that could keep the administration running if Microsoft 365 becomes unavailable.

The honest part: the PoC surfaced technical, operational, and organisational hurdles that complicate broad production rollout — with large video conferences called out specifically. So the program targets 3,000 staff in critical business processes, not the full estate. For context, the federal administration runs more than 50,000 workstations.

Why a Microsoft 365 exit matters for your business

You are not a national government and you do not have CHF 9 million. Take the method, not the migration.

What Switzerland actually bought was information. Before committing, they ran a real workload with real people on the replacement stack and produced a specific list of what holds and what does not. That is the opposite of how most exit conversations go — a spreadsheet comparing seat prices, a vague feeling that leaving would be painful, and no test.

Three things you can copy this quarter, at your scale:

Pick your two highest-risk dependencies. Not everything — the ones where a price change or an outage stops you billing customers. That is usually mail, identity, and file storage.

Run one team on the alternative for two weeks. Not a demo. Actual work. You are not trying to prove the alternative wins; you are trying to produce Switzerland's list — the hurdles, named.

Then price the switch honestly. If the answer is "we stay," you now know that for a reason, and you know your leverage at renewal. If the answer is "the calendar sync would kill us," that is a concrete engineering problem with a cost attached, not a fear.

Note what did not break here: documents, files, mail, calendar. Those are the commodity layer, and the open-source stack has caught up. What broke was large-scale video. That is a narrower dependency than "we're on Microsoft," and narrow dependencies are negotiable.

Sources: Swiss Federal Chancellery, PoC BOSS feasibility study.