All products

CyberSonar24 AI ActAvailable

AI Act compliance without guessing the scope

CyberSonar24 AI Act helps you classify AI systems, assess risks and gather the obligations arising from the regulation, before inspections and fines appear.

01 · Scope

What does CyberSonar24 AI Act cover?

  • Classification of AI systems

    Determining whether a system is in scope and what risk category it has.

  • Organisational obligations

    Mapping AI Act requirements onto roles, processes and documentation.

  • AI register

    One place for an inventory of systems and models in the organisation.

  • Compliance evidence

    Materials ready for an internal and external audit.

02 · Assessments and analyses

Concrete processes in CyberSonar24 AI Act

Each assessment leads from classification and requirements to a documented result and an action plan.

  1. 01

    AIRA: AI Risk Assessment

    A structured AI risk assessment: context of use, people exposed, possible harm, likelihood, severity and mitigating measures.

    Basis: A CyberSonar24 process based on the AI Act, Article 9, and the NIST AI RMF functions: Govern, Map, Measure, Manage

  2. 02

    AI Act classification

    Establishing the organisation’s role and the system category: prohibited practice, high risk, transparency obligation or minimal risk.

    Basis: AI Act, Articles 5-6 and Annexes I and III

  3. 03

    FRIA: fundamental rights impact assessment

    An assessment of the impact of a specific high-risk system deployment on people and groups, together with response measures.

    Basis: AI Act, Article 27. The obligation applies to specified deployers; a FRIA does not replace a DPIA

  4. 04

    High-risk system conformity assessment

    A review of data governance, documentation, logging, human oversight, accuracy, robustness and cybersecurity.

    Basis: AI Act, Articles 8-15

03 · Outcomes

What does the organisation receive?

  • A register of AI systems, models and use cases
  • An AIRA report with risks and a mitigation plan
  • System classification with a rationale
  • A FRIA and a link to a DPIA, where required
  • A map of obligations of the provider, deployer, importer and distributor
  • An evidence pack: human oversight, logs, documentation and monitoring

04

Where the module sits in the platform

CyberSonar24 AI Act runs on the same platform in which the organisation manages compliance with NIS2, the Polish National Cybersecurity System Act (KSC), the GDPR, DORA and ISO/IEC 27001. There is no need to launch a separate system solely to inventory artificial intelligence systems and collect evidence for an inspection. The platform can be run in the cloud or on the client's servers (on-prem).

05

From the register to the documentation

The organisation first records which systems, models and use cases it actually uses. For each entry it establishes its role: provider, deployer, importer or distributor. The same organisation may have different roles in respect of different systems, and the obligations follow from the role. It then records the classification under Regulation (EU) 2024/1689: prohibited practice, high-risk system, transparency obligation or minimal risk, together with a rationale.

The risk assessment (AIRA) organises the context of use, the people exposed, possible harm and mitigating measures, and results in a report with a risk mitigation plan. Where a high-risk system is involved, a fundamental rights impact assessment (FRIA) and a review of the requirements in Articles 8 to 15 can be carried out separately, including human oversight, event logging, data governance, documentation, accuracy, robustness and cybersecurity. If a data protection impact assessment is also required, the link between the two assessments is included in the result. A fundamental rights impact assessment does not replace a data protection impact assessment. Materials from oversight, event logs, documentation and monitoring form an evidence pack ready for an internal and external audit.

06

Competence that the regulation requires of people

AI Act compliance is not only the classification of systems. Article 4 of Regulation (EU) 2024/1689 requires that people who operate artificial intelligence systems have an appropriate level of competence. Two courses are already available in the e-learning module: “Foundations of artificial intelligence at work” and “Safe use of artificial-intelligence tools”. Each course ends with a test that is scored automatically and with a named certificate, and management can see the training level. This evidence can be collected in parallel with work in the systems register, without waiting for the inventory to be completed.

07 · Who it is for

Who is this product for?

  • Providers and users of AI systems
  • Legal and compliance teams
  • Product teams deploying AI
  • Organisations with high-risk models

08

Frequently asked questions

Does the AI Act apply only to providers of artificial intelligence systems?

No. Obligations depend on the organisation’s role in respect of a given system: provider, deployer, importer or distributor. The same organisation may have different roles in respect of different systems. The module records the role against the register entry and shows the resulting obligations.

Does a fundamental rights impact assessment replace a data protection impact assessment?

No. A fundamental rights impact assessment (FRIA) arises from Article 27 of Regulation (EU) 2024/1689 and applies to specified deployers of a high-risk system. A data protection impact assessment (DPIA) arises from the GDPR. Where both assessments are required, their link is taken into account. One does not replace the other.

How can staff competence required by Article 4 of the AI Act be documented?

The e-learning module offers the courses “Foundations of artificial intelligence at work” and “Safe use of artificial-intelligence tools”. Each course ends with a test that is scored automatically and with a named certificate, and management can see the training level. This is evidence that can be shown in an audit.

Does CyberSonar24 AI Act run only in the cloud?

No. The platform can be run in the cloud or on the client's servers (on-prem).