A methodology of observation, not evaluation.
The OMINA methodology defines how an organization can be observed in the information systems used by artificial intelligence: under which conditions, with which limits, and what the results allow — and do not allow — to be concluded.
Organization
The real entity.
Informational representation
Its encoding across information environments.
Reconstruction
The way systems assemble it when they are queried.
Every methodological statement bears on the last two levels exclusively.
These systems do not draw on a memory alone: they also search at the moment of the question. What they return combines what they have retained and what they find — and both can be observed.
OBSERVATION
The infrastructure queries the systems, records what appears — and what does not — and preserves every piece of evidence.
INTERPRETATION
OMINA reads the evidence and establishes a structural reading: what the reconstruction reveals, within the exact limits of what was observed.
GOVERNANCE
The reading allows the organization to understand how it is represented, what builds that representation, and how this may inform its decisions. These orientations are established by OMINA — never generated by the instrument.
Observation executed by the OMINA analytical instrument, under defined protocol conditions; every execution is preserved under evidence custody before being served. Interpretation and governance established by OMINA.
An observation never produces a simple "present/absent". It produces distinct states, never conflated:
OBSERVED
Under the conditions tested, the content appeared when the systems were queried.
MEASURED ABSENCE
Queried, with no appearance — a finding, not a void. The absence was established under defined conditions; it constitutes a result, never a judgment.
NOT MEASURED
These conditions were not tested. No conclusion is drawn from what was not observed.
NO DISTINCT CONTENT
The observation is complete; no distinct content follows for this dimension. The corresponding finding is carried by the overall profile — nothing failed, nothing is missing.
Measured absence and absence of measurement are two different pieces of information. Conflating them would be a category error; the methodology forbids it. A qualified absence is never a judgment of the organization: it describes a state of the information environments, under defined and dated conditions.
An observation is a state, not a trend.
Questions of understanding
Does OMINA correct what AI systems say about an organization?
No. OMINA does not intervene in AI systems and does not modify their responses.
OMINA observes the way AI systems describe an organization — what is retained, what is absent, what is confused — and identifies what is fragile and where it comes from.
OMINA then establishes orientations: what should be addressed as a priority — not what to publish in order to improve how it appears.
Decisions and their implementation remain the organization's own.
Where does what an AI system says about an organization come from?
Not only from its website.
An AI system assembles its answer from the information environments available to it at the moment of the question: what it has retained, and what it finds.
What the answer returns depends on the coherence of these environments with one another — and an organization has no direct way of seeing what they hold about it.
This reconstruction is what OMINA observes and characterizes.
Why should an organization care about the way it is understood by AI systems?
For years, organizations have built their digital presence, presented their activities and their offerings, and explained what sets them apart. But they have no direct view of what AI systems retain from all this when they speak about them.
OMINA provides that reading. Beyond being visible: what does AI actually understand about my organization?
Being visible is one thing. Being understood is another.