How one hundred organizations are reconstructed by AI systems, observed under defined and dated conditions. An observation is a state, not a trend.
Observing, at the scale of one hundred companies, the way AI systems reconstruct an organization, with a protocol, preserved evidence and a structured human interpretation.
Observation
Signals detected
The infrastructure queries the systems, records what appears, and what does not, and preserves every piece of evidence.
Interpretation
Coherence and meaning
OMINA reads the evidence and establishes a structural reading, within the exact limits of what was observed.
Governance
Reconstructed image
The reading allows an organization to understand how it is represented and what builds that representation.
An independent study, built on an explicit protocol and preserved evidence.
The OMINA 2026 · 2027 study observes, analyses and interprets the way 100 Swiss companies are represented, understood and reconstructed by the main conversational artificial intelligence systems, under a documented protocol.
The precise dates will be confirmed by OMINA: this page shows the structure and the sequence, not the final calendar.
2026
One hundred companies selected under a documented protocol
The scope of the study, the unit of analysis and the inclusion and exclusion criteria are defined before any observation. The corpus of 100 companies is built under an explicit, transparent and auditable protocol, ensuring diversity of sector, language, geography and company size.
2026
One common, governed protocol
Each company is observed under the same protocol and in controlled conditions. The interactions, their conditions of observation and the evidence needed for interpretation are preserved. The answers obtained, the absences and the limits of access are documented.
2026 · 2027
An individual and cross-corpus reading
Observations are analysed individually, then compared across the corpus. The analysis examines convergences, divergences, variations and emerging phenomena, without presuming an expected result.
2027First half
Within the limits of the protocol
Results are interpreted in the light of the protocol, the corpus and the methodological limits. OMINA distinguishes what can be established, what cannot, and what remains undetermined.
2027Mid-year horizon
Study and methodological note
The study is published with its full methodological note. The corpus, the observation protocol, the limits of the study and the main results are presented transparently, including where no particular phenomenon is established.
The corpus and the protocol are established before the results are analysed. Companies are not selected on the basis of the answers obtained or of the supposed interest of their case. Results are published within the limits of the corpus and of the conditions of observation, including where no particular phenomenon is established.
Independence
A study conducted without commercial influence and under a protocol defined in advance.
Transparency
The method, the corpus and the limits are documented and made public with the results.
Integrity
Each observation is preserved in its context, including absences and non-responses.
Usefulness
A factual contribution to the debate on how Swiss companies are represented in AI.
« What organizations perceive of their representation by AI systems ». OMINA publishes when an observation is sufficiently established to enter the public record.
“A Lausanne deeptech startup wants to understand how AI systems reconstruct the digital identity of 100 Swiss companies.” September 2026, two pages. The document itself is in French.
You will be informed of OMINA’s publications and of the progress of the study on one hundred companies. Nothing else.