Madrid is the engagement with the fewest moving parts. One country, one supervisory authority, one timezone, one currency, and a train that gets us from our office to yours in about two and a half hours. Everything that consumes the first month of a cross-border project is simply absent, which leaves the actual problem.
Madrid is where Spain's largest institutions have their headquarters: the banks and insurers, the telecoms operators, the energy companies, and the central administration. It is also, less visibly, where a great deal of Spanish-language customer operations for Latin America is run from. The work that comes to us divides fairly neatly into three.
The first is decisioning with an audit trail. A credit file, a claim, a tariff exception, a subsidy application — a person reads it, applies a policy, and the outcome must be explicable to a supervisor and to the affected individual long afterwards. The model is the least interesting component; the evidence trail is the product. See credit decisioning and, for the reporting side, regulatory reporting.
The second is the contact centre. Spanish-language support at Madrid scale is an enormous cost line, and it is also where automation goes wrong most publicly: an agent that is confidently incorrect about a tariff generates complaints faster than it saves salaries. We build these with a hard boundary — retrieval from your own approved material, a refusal path, and escalation that carries the context so the customer does not start again. Customer support agents covers how.
The third is document pipelines in the back office: contracts, invoices, onboarding files, regulatory correspondence. Unglamorous, high volume, and usually the project with the clearest arithmetic behind it.
Across all three, the EU AI Act is now a live design input rather than a future one, particularly for anything touching creditworthiness, employment or access to a public service. Our reading of what that means in practice is on the EU AI Act page.
No. We have one office and it is in Barcelona. Madrid is close enough that an office here would be a pretence rather than a convenience — we take the morning train, do the work in your building, and go home. If you need us on site regularly, we will say up front how often that is realistic.
Yes. There are three in-country regions between the major providers, and open-weight models can be self-hosted inside your own environment where nothing may leave it. For any hosted model we set out its retention and training position at the time of writing, and you should confirm it against the provider's current terms rather than relying on our summary indefinitely.
Sí, con normalidad. Las reuniones, los talleres de discovery y el trato diario pueden ser en español, català o inglés. La documentación técnica y el código se entregan habitualmente en inglés por convención del sector, pero si preferís la documentación funcional en español, se hace así.
As processing like any other, which is the useful part: there is no separate AI regime to learn, only the GDPR and the LOPDGDD applied to a system that happens to use a model. The recurring questions are the lawful basis, minimisation, transparency where a decision is automated, and the right to obtain human intervention. Those become architecture: what is logged, what is retained, and where the person sits. Your DPO owns the assessment.
No. The ranges are published on the pricing page in euros and exclusive of VAT, and they are the same for every city. We do not price by postcode, and being in the same country does not make us cheaper or more expensive — it makes the project shorter, which is the part that actually saves money.
Thirty minutes on the problem first. If it warrants a room and a whiteboard, we will get on the train.
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.
