Document Agents
Understand documents through real language comprehension. Recognition of type, content, and context – not template matching. Every extraction verified through the Decision Layer.
Document Agents in detailOn your infrastructure. Under your control.
Over the past decade Lisbon has moved from tourist secret to tech hub for Western Europe. EDP (Energias de Portugal) and Galp Energia form the energy core, Millennium BCP and Caixa Geral de Depositos the banking spine, Jeronimo Martins (Pingo Doce, Recheio, Biedronka in Poland) is the country’s largest food retailer. NOS and Altice Portugal dominate the telecoms landscape. Add the Portuguese tech champions: Farfetch (the luxury fashion marketplace), OutSystems (the low-code platform with global scale) and Feedzai (AI-driven fraud detection for banks worldwide). The Beato Innovation District and the Hub Criativo do Beato host hundreds of startups - the city pulls talent from across Europe because the cost of living is lower than in Berlin or Amsterdam and the tech ecosystem keeps getting denser.
The first is the CNPD (Comissao Nacional de Proteccao de Dados) together with Lei 58/2019, the Portuguese GDPR implementation law. The CNPD is one of the most consistent data protection authorities in Southern Europe and has imposed several high-profile fines on telecoms operators, banks and public bodies in recent years. The CNPD examines AI systems in particular for transparency, data minimisation and the right to a human decision under Article 22 GDPR. Anyone building AI in Lisbon is building for an authority that takes algorithmic audits seriously.
The second is the EU AI Act, which is directly applicable in Portugal. The Portuguese government has named ANACOM (the national communications authority) and the CNPD as coordinating bodies for the EU AI Act implementation, with the plan to build its own market surveillance authority. High-risk systems - especially in energy and finance - will be systematically reviewed from 2026 onwards.
The third is Banco de Portugal and the CMVM (Comissao do Mercado de Valores Mobiliarios) for all financial services firms. Both authorities have set out joint expectations on AI model governance: explainability, Audit Trail, reproducibility and bias mitigation. Banks such as Millennium BCP and Caixa Geral de Depositos are already working on frameworks of this kind. EU AI Act compliance and financial supervision interlock here.
Millennium BCP uses AI in credit analysis for SME customers and needs models that remain reproducible under CNPD and Banco de Portugal supervision - every refusal has to be justifiable inside the legal time window. EDP operates smart meter infrastructure in Portugal, Spain and Brazil and uses AI for load forecasting and anomaly detection in the grid - the models have to be documented for the energy regulator ERSE. Feedzai develops fraud detection models for banks worldwide, headquartered in Lisbon - every model update has to remain traceable in an Audit Trail because the customers run in regulated markets. Farfetch uses recommendation algorithms for luxury products in more than 190 countries - here CNPD requirements collide with the data protection regimes of the UK, the US and China.
In every scenario the question is about AI models that do not live in the lab but in regulated business logic with measurable consequences. The Decision Layer addresses the architecture question: Audit Trail down to SQL level, Human-in-the-Loop on escalation-relevant cases, Cert-Ready by Design.
A Portuguese particularity is the Comissao de Trabalhadores - the Workers’ Committee with information and consultation rights when AI systems are introduced. We design the Decision Layer so that these duties do not have to be retrofitted but live inside the governance. That is particularly relevant for groups like Jeronimo Martins, whose headcount is in the five-figure range and whose AI decisions have direct effects on staffing and shift planning.
Gosign has its own office in Lisbon (gosign.pt) with a Portuguese team - native speakers, local project management, direct contact to the CNPD, Banco de Portugal and CMVM. The Lisbon office is the regional hub for the Portuguese market: discovery workshops, sprint reviews and steering happen on site at the client in the Beato Innovation District, in the Hub Criativo do Beato or directly at the corporate headquarters of Galp Energia, EDP, Millennium BCP, Caixa Geral de Depositos and Jeronimo Martins. For the Portuguese tech champions Feedzai, OutSystems and Farfetch we work as peers in the local engineering culture, with workshops and code reviews in Portuguese, compliance documentation in Portuguese and English, and the Comissao de Trabalhadores at the table from the start. We cooperate closely with Startup Lisboa and the local talent pool. For LATAM projects the Lisbon team additionally pairs with the Sao Paulo office (gosign.com.br) - Brazil and Portugal as one connected lusophone market with a shared language, a shared business culture and an architecture that serves LGPD and GDPR in parallel. The Hamburg headquarters takes care of architectural questions of principle and the link to German and Northern European parent companies.
Portugal is small enough to bring several industry stakeholders directly into the same AI project, and large enough to carry real compliance demands. Anyone who builds a Cert-Ready by Design pilot in Lisbon has a reference case that has passed under CNPD and EU AI Act standards and is then defensible in Brazil (LGPD), Spain (LOPDGDD) and the rest of Europe. The Lisbon Tech Hub, Startup Lisboa and the Beato Innovation District supply talent, investor access and pilot partners.
We bring the experience of building AI systems that stay productive even under strict CNPD supervision - with Governance by Design and the discipline shaped by German data protection practice. Lisbon is the right market in which to scale that discipline across the Iberian peninsula. And for enterprises that use the Portuguese market as a springboard into the lusophone space (Brazil, Angola, Mozambique), Lisbon is the only European location from which that bridge can be built architecturally cleanly.
Not because of technology – but because of missing governance. Without clear rules defining who makes which decision, every AI agent stays a pilot project.
That is why we build every agent exclusively with a Decision Layer. It breaks down every business process into individual decision steps and defines for each step: human, rule engine, or AI. No agent goes into production without this layer.
Understand documents through real language comprehension. Recognition of type, content, and context – not template matching. Every extraction verified through the Decision Layer.
Document Agents in detailSteer business processes across multiple systems and decision points. One agent, complete orchestration. Every step in the audit trail.
HR AI AgentsAnswer questions from enterprise knowledge – with source reference, rule version, and validity date. No verified source, no answer.
Knowledge Agents in detailAuditable. Compliant. Enterprise-grade.
Human-in-the-Loop architecturally enforced – not optional
Complete audit trail for every agent decision
GDPR compliant by design – all data on your infrastructure
Works council compatible – agreements as constraints in the Decision Layer
EU AI Act compliant by design – transparency, explainability, human oversight
Model-agnostic – no vendor lock-in, you own the source code
1 week
Process analysis, understand rule sets, prioritise use cases.
3–4 weeks
Productive PoC. One agent, one process, live on your infrastructure.
Continuous
More agents, more processes. Same governance, same auditability.
After 12–18 months, you operate your agents independently. Source code, prompts, and rule sets are yours.
Analysis and insights on enterprise AI, governance, and agent architecture.
Most AI projects fail not because of technology but because nobody defined the rules. Why the operating model matters more than the language model.
The EU AI Act directly affects HR processes. Risk classification, bias monitoring, human oversight - what is now mandatory and how to prepare.
Agent governance is not an IT topic. It's an HR leadership topic. What CHROs need to know before AI agents enter core HR processes.
“Even as a global market leader, you want to keep moving forward. It is reassuring to have the technological expertise and infrastructure experience of Gosign on our side.”
Head of Innovation, Sony Music Entertainment
“Gosign is not just about speed. It's about how much essential work happens in this time.”
Head of Customer Service & Technical Support, Libri GmbH
Yes. Our Lisbon office (gosign.pt) provides local project management and client support in Portugal.
GDPR compliant by design. All data remains on your infrastructure. No data transfer to third parties. The EU AI Act is directly applicable in Portugal as an EU member state.
CNPD (National Data Protection Commission) supervises GDPR in Portugal. Our agents meet the transparency, explainability, and human oversight requirements.
Yes. In Portugal, the Workers' Committee (Comissao de Trabalhadores) holds information and consultation rights for the introduction of AI systems. The Decision Layer with Human-in-the-Loop architecturally enforces human review for decisions affecting workers.
Talk to us about a specific use case in your organisation.
Schedule a consultation