1. What AI does in Granex
AI assists with support tasks, always under human supervision:
- Drafting of briefs, summaries and translations.
- Summarising and comparing documents and case files.
- Answering questions about immigration rules and procedures.
- Analysing and classifying legal updates (official gazette, case law).
2. Model used
AI features rely on Anthropic's Claude models (claude-sonnet-4-5 family). We do not train our own models on your data, nor do we build profiles from it.
3. The lawyer always decides
AI never makes legal decisions or files anything on its own. Every output is a draft that the lawyer must review, verify and validate before use. Professional responsibility rests entirely with the lawyer.
4. Your data does not train models
Content you send to AI features is used only to generate that session's response. It is not used to train or improve third-party models. Anthropic processes the data as a Processor, under Standard Contractual Clauses and with no retention for training.
5. Limitations and "hallucinations"
Language models can produce incorrect, outdated or fabricated information (including non-existent statutory or case-law references). They do not replace legal judgement. Always verify citations, deadlines and legal grounds against official sources before filing anything.
6. Human oversight and control
You can use Granex without the AI features. Every AI action is recorded in the account's usage, and the firm's administrator can review consumption from the billing panel.
7. Regulatory framework (EU AI Act)
Granex does not provide high-risk AI systems within the meaning of Regulation (EU) 2024/1689 (AI Act): the features are assistance and content-generation under human control. We meet the applicable transparency obligations by informing you that you are interacting with an AI system.
8. Contact
For questions about AI use or to object to automated processing, write to dpo@granex.es.