Dynamic questionnaires
Questions tailored to your application
Questionnaires adapt to your answers and stop as soon as the outcome is settled.
Answer three questions to determine whether your application is an AI system under the AI Act.
Fewer questions, better answers
Every questionnaire is a decision tree. This has three consequences:
Only relevant questions
Follow-up questions depend on the answers given. Parts of legislation that do not apply are never shown to you. Smart cross-references between legal provisions save time, e.g., human oversight requirements under GDPR and the AI Act.
The answers are the evidence
The route you take through a questionnaire is recorded. The result is therefore not just a classification, but also the reasoning behind how that classification was reached.
Every question refers to a legal article
Every question includes a reference to the relevant legal articles. This provides the overview needed to stay in control of compliant AI.
Is your application an AI system?
Answer the first question and the next one appears. Or it does not, if your answer has already determined the outcome.
- Q1 · Outcomecurrent question
- Q2 · Designnot reached yet
- Q3 · Rulesnot reached yet
Outcome
Answer all questions first.
What is the outcome of the application?
One feeds the next
First complete the Identification questionnaire (does the application fall under the AI Act, the GDPR etc.). Then complete the follow-up questionnaires.
- 1
Identification
Is the application an AI system? Does it fall under the GDPR?
Passes onAI system: yes - 2
Role and status
Is your organisation a provider, deployer, importer, distributor or authorised representative? Is the system in use or still in development?
Passes onRole: deployer - 3
Risk category
Is the AI system prohibited or high-risk? Or do transparency obligations apply?
Passes onCategory: high-risk - 4
Obligations
An overview of the obligations that must be met, including templates to help fulfil them.
Passes onOverview of obligations. Templates help to fulfil them.
Questions about the questionnaires
The dynamic questionnaires were developed and tested by Algorithm Audit – European knowledge platform for responsible AI. The questionnaires are publicly available and subject to community feedback. This keeps them accurate and up to date.
Leave the question open. The questionnaire stops at the last question you answered and the application stays marked as unclassified, so an incomplete assessment never looks finished. Questions can be assigned to colleagues who have the expertise to answer them.
Yes. Changing an answer discards the answers that depended on it. If the change affects a later questionnaire, that questionnaire must be completed again. Earlier completed versions of the questionnaires are stored.
No. The questionnaires help you build a defensible justification of why applications do or do not fall under certain legal requirements, and how those requirements can be met. Whether that classification is correct for your application remains the responsibility of your own organisation.
Prepare for the 2 December 2027 high-risk AI Act deadline
Start using ClassifymyAI today.
or send an email to beta at classifymyai.com