Skip to main content

What is QA

QA ZERO is one edition of something larger. Understanding that larger thing — QA, the trunk the editions grow from — is the fastest way to understand why the API is shaped the way it is.

QA is a Behavioral Intelligence Architecture​

Traditional web analytics reports metrics: page views, sessions, bounce rate, conversions. QA has a different objective.

QA exists to help AI understand how humans interact with information.

Rather than producing reports for humans to read, QA builds structured behavioral knowledge that AI can reason with. That is why QA is not, at its core, an analytics platform. It is a Behavioral Intelligence Architecture.

Raw behavioral events carry no meaning on their own. QA converts observable behavior into structured knowledge, and the goal is not collecting data — it is enabling AI to understand human intent.

Human → Behavior → Structured Events → Semantic Labels
     → Knowledge → Reasoning → Recommendation → Better Human Experience

Visualization is only one possible output. Understanding is the real objective.

Three principles for every behavioral event​

  • Observable — objectively measurable (scroll, click, hover, reading time, revisit, navigation).
  • Contextual — the same behavior means different things on different sites. A three-minute read means engaged learning on documentation, hesitation on an e-commerce page, and normal on news. Context is therefore part of the data, not metadata around it.
  • Semantic — meaning should not rely on statistical inference alone. QA supports explicit semantic labels (reading carefully, comparing products, hesitating, task completed), which become high-quality signal for AI.

Why the labels are accurate: collective intelligence​

Labels are only valuable if they are accurate, and accuracy is the hard part — because the same behavioral value means different things depending on the type of website. So labels must be segmented by context (site type), not applied globally.

QA's answer is collective intelligence: across a distributed network of WordPress sites, humans and AI collaborate to raise the inference accuracy specific to each site segment. This mechanism is the subject of patent JP 7011367. Because the data stays on each site's own server — Cookie-less, anonymized, no central cloud — QA is a distributed, first-party counter to centralized analytics: each site runs autonomously, yet all of them improve the shared understanding.

How this shapes QA ZERO's API​

Everything in this developer manual follows from the ideas above:

  • QAL, not SQL. QAL is a small declarative language shaped so that an AI assistant composes a correct query on the first try, and so that a malformed query cannot damage the backing store.
  • Materials carry meaning, not just rows. The materials you query already encode behavioral context (deep-read, irritation clicks, engagement depth), not just raw hits.
  • The spec is machine-readable and first-class. The AI-facing spec ships next to the human docs so that LLMs, MCP servers, and agents consume it directly.

QA does not replace GA4​

QA is meant to be used alongside GA4. GA4 measures what happened; QA adds what it meant and what to do next, so that people can act on their own data instead of decoding numbers. The real purpose is a creativity loop: when people can finally use the power of their data, they spend less energy interpreting metrics and more energy expressing their own creativity — and that creativity, observed as behavior, feeds back to make the intelligence better.

Where to go next​