Copenhagen Worked from Barcelona Documents and sensor data Same hour as us

Copenhagen.
Documents and sensors.

We are based in Barcelona and work on your hours exactly. Copenhagen is unusual among European cities in that the two hardest kinds of industrial data are equally present here: paper that moves goods across borders, and telemetry from assets standing in the North Sea. They are not the same problem and they should not be the same system.

Last reviewed 5 September 2026 · Engineering guidance, not legal advice.

WHAT IS ACTUALLY DIFFERENT HERE

Two data shapes, one city

A Copenhagen group will often own both a document estate — bills of lading, customs declarations, certificates, regulatory dossiers — and a time-series estate coming off vessels, turbines and production lines. Language models are good at the first and largely irrelevant to the second, where the useful work is anomaly detection and forecasting with a language layer only at the point a human reads the result. Selling one platform for both is the most common way this goes wrong, and we will argue against it even when it costs us scope.

No Danish region, and it barely matters

At the time of writing the major providers have no Danish cloud region: the nearest are in Sweden and Germany. Both are EU, so jurisdictionally nothing changes, and the latency across the Øresund is not a problem for anything we would build. It is worth checking whether your own contracts say Denmark or say the EU, because those are different requirements and only one of them is expensive. If it genuinely says Denmark, the route is self-hosting on Danish infrastructure and we will price that honestly.

Shipping data is cross-border by definition

Datatilsynet supervises data protection under the standard GDPR, which is the simple part. The complication in this city is structural: a single shipment involves shippers, forwarders, customs authorities, ports and consignees across several jurisdictions, some well outside the EEA, and the documents carry named individuals. The transfer question here is not theoretical and it is not solved by a region setting. It is solved by deciding what leaves your perimeter, which is a design decision made early or a problem discovered late.

Life-science documentation has its own gravity

Danish pharmaceutical and biotech work answers to documentation standards that were strict long before anyone here had an AI question. What that means practically is that a summary without a traceable citation is not a lesser deliverable, it is an unusable one. The requirement is retrieval you can follow back to the source paragraph, and that is a design constraint we would rather accept at the start than negotiate at validation.

THE WORK THAT COMES OUT OF THIS CITY

Shipping and logistics is the sector that defines Copenhagen commercially, and it produces the most reliably valuable document work in Europe. A freight operation handles enormous volumes of semi-structured paper that arrives in every possible condition — emailed, scanned crookedly, photographed, in several languages, occasionally handwritten. Extracting it correctly is worth real money per document, and the engineering that decides success is the exception path rather than the happy path. Our logistics document pipeline is that pattern written down, and the wider sector view sits under logistics.

Pharmaceuticals and biotech arrive with the opposite tolerance for error. The valuable systems here read across regulatory submissions, study reports and internal literature and answer a specific question with the paragraph it came from, so a scientist can verify it in seconds. Fluency is not the requirement; verifiability is. That is what we build under life sciences research.

Renewables and wind bring the sensor half of the city. Turbines generate continuous condition data, and the commercial question — should this asset be serviced now or in six weeks — is a forecasting problem with a maintenance-history document problem attached to it. The two halves are usually owned by different teams and stored in different systems, which is most of why this work is hard. The approach is described under predictive maintenance.

Design and consumer brands bring a smaller but cleaner problem: large product catalogues sold across many markets, where search, translation quality and merchandising decide conversion. It is the least regulated work we do in the city and often the fastest to show a result.

HOW A COPENHAGEN PROJECT ACTUALLY RUNS

We come up for discovery, often twice over

Barcelona to Copenhagen is a direct flight of about two and a half hours. Where both halves of the estate are in scope, the document people and the asset people are usually in different buildings and sometimes different towns, and the first workshop has to reach both — so we plan discovery around two sites rather than one meeting room. On a normal engagement we do not bill the travel as a line item.

Then remote, same hours all year

Denmark and Spain keep the same clock, including the daylight-saving switch, so a question asked in a morning standup is answered the same morning. Nothing about the schedule needs converting.

And back for handover to the people on the line

The ownership session goes to your engineers, and for a document pipeline we also run it with the operations team who will see the exceptions, because they are the ones who decide whether the system is trusted. After that it is yours, in your repositories and your cloud accounts.

QUESTIONS

QUESTIONS — 5

No. We have one office and it is in Barcelona. We travel to Copenhagen for discovery, for handover and for anything where being in the room changes the outcome. There is no Danish entity and no Nordic partner behind us.

In the EU, straightforwardly. In Denmark specifically, not with a major provider region at the time of writing — the nearest are in Sweden and Germany, both EU, which satisfies most requirements that are actually written down. Where a policy genuinely names Denmark, the route is a self-hosted open-weight model on Danish infrastructure, and that is a larger build we would price rather than approximate. Where a hosted model is used we document its retention and training terms at the time of writing, and you should confirm them against the provider's current terms.

Yes, but not with the same tools, and this is the answer we most often have to defend. Condition monitoring and forecasting on time-series data is classical machine learning and signal work; a language model belongs at the end of it, turning a result into something a planner can read, and sometimes not at all. If you have been quoted a single AI platform covering both, ask what model is reading the vibration data and what it was trained on.

It is a question rather than a problem, and it needs answering in the architecture rather than in the contract alone. In practice we design so the extraction runs inside your perimeter and only the structured result — often with the personal data already dropped, because a customs line item does not need a name — moves onward. That removes most of the transfer surface without changing what the business gets. Your DPO owns the legal analysis; our job is to make it a short one.

The same as anywhere else — the ranges are on the pricing page, in euros and exclusive of VAT. We do not price by postcode. A project covering both the document estate and the asset estate is two systems and is estimated as two.

Keep reading

Architecture

Logistics Document Pipeline

→

Solution

Predictive Maintenance

30-minute technical call · no deck