Expert insights on AI traffic, content monetization, and GDPR compliance for publishers.
Before any control of ours starts rejecting requests, we prove in a log-only mode that it is complete, that it actually fires, and that it never trips legitimate traffic.
Two changes make your traffic intelligence more trustworthy: traffic over your plan limit is still measured instead of going dark, and geography is read from the real visitor rather than the delivery edge.
Before any destructive change to our compliance trail database, an automated verification pass confirms the operation will not lose audit-trail integrity. The pattern caught a real divergence this session and prevented data loss.
When privacy notices, marketing pages, PDF DPAs, and the live service all reference the same retention periods and processing limits, drift between them is a real misleading-claim risk under EU consumer-protection law. Here is how HumanKey closes that class of drift mechanically.
When plan tiers and feature gates spread across many code paths, drift is inevitable — unless you mechanically enforce a single source of truth. Here's how HumanKey closes that class of bug across our customer-facing surfaces.
Three analytics tools, three different visitor counts on the same site. We explain what each one measures, why the gap between them is structural, and which number to use for which decision.
An operator-side update on the review pattern we use to keep our customer-facing surfaces consistent. We publish the methodology — not the findings — and explain why.
What we learned after 90 days of running an AI-transparency banner under EU AI Act Article 50: what we measured, where the wording was harder than the engineering, and what we'd change.