ActBrief uses a deterministic rules engine to turn product facts into sourced risk hypotheses, missing facts, documentation gaps, and questions for counsel. It does not issue a legal classification, certification, or conformity assessment. The methodology was reviewed on 27 August 2026; external legal review remains pending.
ActBrief is a preliminary technical readiness tool, not legal advice. It produces hypotheses from submitted facts and a rules engine. It does not certify compliance, perform a conformity assessment, or replace product-specific review by qualified counsel.
The scanner maps intended purpose, provider or deployer role, territorial signals, affected persons, material influence, profiling, automation, model provider, and data categories to potential AI Act issues. It produces risk signals, documentation gaps, effective dates, missing facts, and questions for counsel.
Dated content includes a last-reviewed date and, where relevant, a change log. Material changes to application dates, Commission guidance, or scope interpretation should update the relevant source trail, checklist, template, and scanner output together.
Written by: ActBrief Editorial (founder / editor).
Legally reviewed by: not yet — we do not invent an “ActBrief Legal Team.” When an EU tech / regulatory lawyer reviews the methodology and named pillar pages, we will publish their name, credentials, review date, and scope of review (for example Article 50 interpretation and effective dates).
Until then, labels stay explicit: external legal review pending. Pillar pages that matter most for review once counsel is engaged: this methodology, /checklist, /article-50, chatbot disclosure, chatbots citation page, and OpenAI SaaS role.
Free reports are generated by a deterministic rules engine from the information submitted. Product descriptions are not sent to a third-party language model to generate the report. See the Privacy Policy for retention and processor details.