Large venues (exhibitions · museums)Illustrative scenariosAudience · Technical decision-makers (CTO · CISO)
Reading flow without faces
Six months of operation atop anonymization. A record of returning underperforming spots back to translation.
From problem to result
- Problem
- Video flow analysis shelved over privacy and security concerns
- Applied and measured
- Flow analysis in exhibition spaces was thought impossible without facial recognition. We ran flow without faces for six months atop Anonymizer and Spatial AI, and at re-measurement, honestly returned underperforming spots back to the translation stage.
- Expected output
- Flow operated atop real-time anonymization · effectiveness verified by re-measurement
Anonymizer (SEAL) · Spatial AI (MTMC) · saai insight
Illustrative
Illustrative: 18 of 22 (about 82%)
Spots proven effective
Illustrative
Illustrative: 4 → back to translation (stage 3)
Underperforming spots
Illustrative
0 cases (guaranteed by SEAL anonymization)
Facial recognition
Illustrative
Illustrative: about 94% accuracy
MTMC tracking
Evidence and scope
05 Re-measureIllustrative scenariosAudience · Technical decision-makers (CTO · CISO)
We can see flow without faces.
It’s data that doesn’t hide what underperforms.
The only non-retail case of the five. The archetype of the re-measurement stage — “we don’t hide what underperforms” — and the first proof of Beyond Retail.