K-anonymity is the privacy rule that protects small data segments from disclosure. On Agentics it means a benchmark segment is never displayed unless it contains at least k = 5 distinct organizations.
Aggregate statistics can leak confidential information when a slice of data is small enough to point at a single customer. To prevent this, any segment covering fewer than five organizations is suppressed entirely — it is not shown, estimated, or rounded into view. Because the floor is enforced before any benchmark snapshot is published, an individual organization's performance can never be reverse-engineered from a public chart. Confidentiality is structural, not a setting that can be switched off.