Services

MIRAI offers two complementary services: a modular software toolbox for AI auditing, and consulting engagements for organizations that need structured support throughout the process.

MIRAI

Toolbox

MIRAI Toolbox is a suite of modules for the analysis and verification of AI systems.

It is designed to be customizable: organizations can adopt individual modules based on their specific needs, or deploy the full suite.

  • Bias Detection: Identifies and measures bias in classification and scoring models across protected attributes, geographic areas, and organizational categories.

  • Risk Assessment: Quantifies risk exposure through the r-index and isorisk score, three-dimensional metrics developed from MIRAI’s published research.

  • Opacity Analysis: Evaluates the explainability of AI models, flagging components that do not meet transparency requirements.

  • Dashboard: Centralizes audit results in a structured, readable interface for technical and non-technical stakeholders.

  • Automated Reports: Generates standardized documentation suitable for regulatory submissions and internal governance processes.

  • GenAI Module (in testing): An extension of the Toolbox for the analysis of generative AI systems. Currently in testing with selected partners.

MIRAI

Consulting

For organizations that need support beyond the toolbox, MIRAI offers structured consulting engagements.

Collect information

about the AI models and data currently in use.

Identify and prioritize

the systems and scenarios that require assessment.

Audit data and models

using MIRAI tools and methodologies.

Deliver actionable metrics

and a mitigation roadmap.

Support implementation

of the roadmap toward responsible AI adoption.

Training

We also offer standalone training sessions for technical and non-technical teams on responsible AI practices, audit methodologies, and EU AI Act compliance.

Subscribe to our newsletter

Reasponsible AI updates directly to your inbox.