Your answers stay in your browser
The classification runs locally. Only the identification details you fill in at the end are recorded.
Certification · free tool
Nine questions, five minutes. At the end you will know which risk tier the AI Act places your system in, which obligations are enforceable against you, which deadlines are running, and what documentation you will have to produce.
The transparency obligations of Article 50 become enforceable on 2 August 2026 against any organisation placing AI-generated content on the European market. The Commission published guidance on 20 July; ANACOM still has no enforcement unit in place.
We know of no other ethical AI assessment tool in European Portuguese, built on the text of the regulation and offered free of charge. If you know of one, tell us and we will publish the link — the goal is for the country to comply, not for ENIA to be right.
The classification runs locally. Only the identification details you fill in at the end are recorded.
It is an indicative self-assessment based on the text of the regulation. The final classification is the operator’s responsibility.
ENIA does not provide compliance services. If you need help, the meeting is free and does not end in a pitch.
Last step
These details serve a concrete purpose: nobody knows the real state of AI in Portugal, because nobody measures it. This test is the first instrument capable of measuring it — and it only works if we know which sector and which company size each answer comes from.
INE measures whether companies use AI. Nobody measures whether that AI is classified, documented and governed. It is the most consequential gap in the national diagnosis: without it, any AI policy is made blind.
Every completed test becomes a point on a public chart — risk distribution by sector and by company size. We publish the aggregate, never the individual case, unless expressly authorised above. The data is not shared with third parties and not used to sell anything, because we sell nothing.