Hamburg runs on paperwork and on weather. Germany's largest port, a publishing industry with archives going back generations, aircraft manufacturing and an offshore wind sector all produce the same underlying problem in different clothing: far too many documents and readings, arriving continuously, mostly routine, occasionally very much not.
Freight and port logistics is the largest single source of work here, and it is more interesting than it sounds. A shipment generates a chain of documents produced by different parties in different formats and different languages, each of which has to agree with the others before goods move. Most of the labour is reconciliation: reading a document, finding the corresponding entry somewhere else, and deciding whether a discrepancy is a typographical difference or a real problem. That is a good fit for a system precisely because a human doing it at three in the morning is not at their best, and it is a dangerous fit unless the system is honest about its own uncertainty. Our logistics document pipeline describes the shape we use, with the exception path as the main feature rather than an afterthought.
Publishing and media groups arrive from the opposite direction. Their problem is not throughput but recall — decades of articles, images and broadcast material where the person who knew how it was filed has retired. Retrieval systems here live or die on the metadata work nobody wants to fund, and on being able to show where an answer came from. The architecture is covered under retrieval systems.
Aviation manufacturing and the maintenance businesses around it bring configuration and compliance records with a long life and a hard traceability requirement, which is a document problem with an engineering accent.
Renewables, and offshore wind in particular, bring sensor histories, service reports and weather data that only become useful when joined. Failure prediction on rotating equipment is a genuinely well-suited application, and it is described under predictive maintenance — with the same caveat that reaches every asset-heavy project: the model is easy and the data plumbing is not.
No. We have one office and it is in Barcelona. We come to Hamburg for discovery and handover and for the site visits that make freight work comprehensible. There is no local team and we are not going to imply one.
Usually, and the honest answer starts with a question rather than a yes: what does it expose. Some of these systems have modern interfaces, some have a database you are allowed to read, and some have a nightly file drop that everyone pretends is an integration. All three are workable. Which one you have decides most of the timeline, which is why we look at it during discovery rather than after signing.
Less accurate than people assume, provided the system knows when it is unsure. A pipeline that handles the routine majority and escalates the rest with a clear reason removes most of the labour and adds no risk. A pipeline that is slightly better on average but guesses silently on the hard cases is the one that causes an incident. We build for the first and measure it against your own historical documents, not against a public benchmark.
That depends entirely on what gets logged, and it is a decision rather than an accident. We scope logging to what the system needs to work and to what an auditor would need to reconstruct a decision, and we write down what is not collected. Where a works agreement is needed, having that description in advance makes the conversation shorter. Your HR and legal functions own the agreement itself.
Our ranges are on the pricing page, in euros and exclusive of VAT, and they do not vary by city. Document pipelines are one of the few things we can scope tightly early, because the volume and the document types are usually known before anyone talks to us.
Thirty minutes on the volumes, the exceptions and what a pipeline would realistically remove.
30-minute technical call · no deck
What we built, what stage it is at, and a link to the live thing wherever there is one.
yourOS is an operating system with an AI at its heart. You talk to it like a person, out loud or typed. It makes your calls, plans your days and works on your projects with a crew of agents, on your own private server. Our own product, built from the first commit; the waitlist is open.
Website, backend and app assets for Sonora, a free AI sound-therapy app on iOS and Android published by Cognitive AI.
Brand, website, prototype, apps on five platforms, the AI backend and the algorithm-intelligence pipeline for Podswap, a social growth engine that runs without a human in the loop.
In closed beta: Roxy, a private AI assistant with permanent memory running on infrastructure the user owns. Brand, site, product, backend and conversational design.
Live at coherence.com: brand identity generated with AI agents we directed, the motion system and website for the group behind Sonora, AboutAI, WakeUp and MudShadows. Agent work remains in build.
Live at domainbooth.com for DomainBooth, a premium domain brokerage: a new animated mark, every page redesigned and rebuilt with designed mobile layouts, and a custom CMS for its sales record, featured names, press and enquiries.
Apps, platforms and campaigns for these teams, long before the AI work. Every one is written up on the work page — what we did, and where our part ended.
