Program · the evaluation
A single evaluation against the four reference frameworks. You come out with an AI Index, a profile across six properties, and a roadmap. Every number reproduces from the inputs.
In short: the AIiD program (programme iDIA) is identifiable's evaluation program. It evaluates every AI system and agent against the six properties of the AI id framework, on evidence, and produces the AI Index (0 to 100), the profile, and the roadmap. At the threshold (an Index of at least 70, no property below 60), it opens the Responsible AI Practice designation.
Three moves. Each one produces a deliverable that the next one uses.
01
Map every AI system in use: intended uses, data touched, dependencies, supervision in place. Most organizations discover here uses that no one had named.
02
Evaluate each system against the six properties of the AI id framework, on evidence. Each score is tied to a level of evidence: Absent, Declared, Documented, Operational, Measured and improved.
03
Calculate the AI Index, produce the profile and the roadmap. At the threshold, grant the "Responsible AI Practice" designation. Below the threshold, target the gaps and plan the re-evaluation.
A score from 0 to 100, the weighted sum of the six properties. The weighting is derived from the four reference frameworks via an emphasis matrix: every weight is justified before an audit committee.
Where averages hide, the profile shows. Each property carries its score and its evidence level. It is the profile, property by property, that determines the designation.
Gaps, prioritized by their contribution to the AI Index. You know what to fix first, and what each correction returns to the score.
The "AIiD program · completed" certificate attests the process. The "Responsible AI Practice" designation is granted only at the threshold: an AI Index of at least 70, with no property below 60.
The AIiD program looks where the exposure actually lives: in the systems, in the agents, and in day-to-day usage.
An agent triggers actions on its own. The evaluation documents its access, its autonomy, its guardrails, and what it can do in the organization's name without a human seeing it happen.
What level of autonomy is acceptable, on which decisions, with which data. Tolerance is decided with leadership, in writing, and the evaluation measures every system against it.
Confidentiality and privacy in the flows; business, data, and reputation exposure. The profile shows where an incident would cost the most, and the roadmap fortifies those points first.
Duration depends on the number of systems in use and the availability of evidence. Scope is defined together at the scoping stage. The most useful first step is a thirty-minute discovery call, or the AI Index diagnostic online.
AIiD is the evaluation program. At its end, identifiable issues the « AIiD program · completed » certificate and, at the threshold, certifies your organization's conformity in AI governance, the NIST AI RMF, and Loi 25, then accredits the « Responsable AI company » designation (« RAI company » mark). For ISO/IEC 42001, the cabinet prepares your organization for certification issued by an accredited body.
An AI Index from 0 to 100, a detailed profile across six properties, a roadmap prioritized by gap, a program completion certificate, and at the threshold, the "Responsible AI Practice" designation.
All the better. Most consultants deploy. identifiable makes what is deployed measurable, traceable, and accountable. Your consultant builds the system; identifiable produces the score your board will ask for.
Read next: the AI id framework, the six properties in detail · the Responsible AI Practice designation · the AI Index diagnostic