Entit
Get in touch
All insights
Data sovereignty 2 Nov 2026 · 4 min read

Denmark's shelved Office pilot: sovereignty needs results, not just first steps

In June 2025, Denmark's Ministry of Digitalisation announced what the minister called a first step towards digital independence. A group of staff in the ministry would use Collabora, an open-source office suite based on LibreOffice, instead of Microsoft Office – built into F2, the ministry's case-management system.

In October 2026, the Danish engineering weekly Ingeniøren reported that the project has been quietly shelved, without a proper evaluation. According to Ingeniøren, 38 employees took part. The planned evaluation never happened.

It would be easy to read this as proof that leaving Microsoft does not work. That is not what it shows. It shows something more useful for anyone in Swedish public sector or healthcare planning their own sovereignty work.

The test criteria were the right ones

The ministry's own announcement listed what it would assess: whether Collabora worked with the ministry's templates and the format for government cases, whether track changes and tables behaved, and whether documents could move to and from Word without changing their layout.

That is a good list. It recognises that an office suite is not a standalone product. It sits inside a web of templates, case systems, exchange with other authorities and decades of documents. Most of the cost of a switch lives there, not in the licence.

But a good list of questions is worth little if nobody writes down the answers.

An unevaluated pilot teaches nobody anything

A pilot exists to reduce uncertainty. When it ends without a result, the uncertainty stays where it was – and the next organisation has to pay to learn the same lesson again.

The Danish case also shows how fragile pilots are when they depend on one sponsor. Ingeniøren notes that the project was launched by a minister in the previous government. When priorities shift, a project with no agreed exit criteria and no published findings is easy to let fade.

Meanwhile, a different Danish track continues. Statens IT, which provides workplace IT to 166 client organisations, is piloting SIA Open, an open-source version of the state's standard workplace. It started small at the Road Safety Authority, built the central management it needed from scratch, and plans a release that client organisations can order from February 2027. Same country, same goal, different method: a service with an owner, a roadmap and a delivery organisation behind it.

What we would do differently

Based on our work with architecture and sovereignty in Swedish organisations, this is how we would frame a pilot like this:

  • Choose the workload by risk, not symbolism. Office documents are visible, but the sovereignty question that matters most may be in your case system, your identity provider or where your logs end up. Start where dependency and data sensitivity are highest – or where a move is cheapest to prove.
  • Map the integrations first. Templates, case and records systems, e-signing, archive formats, and the documents you exchange with others. That map is the real scope.
  • Write success and exit criteria before day one. What result means "scale up", what means "adjust", what means "stop"? Agree them with the business owner, not just IT.
  • Measure, then publish. Even a "not yet" is valuable – for your own organisation and for others facing the same choice. In the public sector, sharing results is part of the return on the investment.
  • Give it an owner beyond one sponsor. A pilot that only lives in a minister's or a CIO's agenda dies with that agenda. Put it in a delivery organisation's plan, with budget for change work and support, not only licences.
  • Decide per workload, not all-or-nothing. Some workloads can move now, some later, some may stay. A written decision with a date beats both a big-bang switch and a quiet drift.

Sovereignty is a programme, not a pilot

The interesting question from Denmark is not whether Collabora can replace Word. It is how an organisation learns, step by step, which dependencies it can reduce, at what cost, and in what order.

That learning only compounds if each step ends with a written result. Otherwise every pilot starts from zero.

Sources

Discuss this with us
An unhandled error has occurred. Reload 🗙

Rejoining the server...

Rejoin failed... trying again in seconds.

Failed to rejoin.
Please retry or reload the page.

The session has been paused by the server.

Failed to resume the session.
Please reload the page.