Search is being replaced by selection. The framework exists to make selection measurable.
For two decades the question was ranking: where a page appeared in a list of results a person would scan. AI search changes the question. A system retrieves a handful of sources, reads them, and cites a few. There is no page two. An organisation is either selected as a citable source or it is invisible, and the difference is decided by signals most businesses have never looked at.
Those signals existed before the framework, scattered and unnamed: whether crawlers can reach a site, whether its content is machine-readable, whether its identity and authorship are legible, whether it is structured to be trusted. What was missing was a model that organised them into a coherent whole, a way to talk about digital authority as something observable rather than something asserted.
The framework borrows its organising idea from the periodic table: position and grouping carry meaning. Signals are arranged into related families, with primary signals that most directly affect selection, supporting signals that qualify them, and emerging signals that are becoming relevant as AI systems evolve. Arranging them this way turns a checklist into a model, one that can be measured, researched and extended.
The framework is deliberately stable and slow-moving, because its job is to be the constant against which change in the world is measured. A framework that shifted underfoot would make longitudinal research meaningless. This is what lets the research programme at ptoda.org publish findings that remain comparable across years.
The framework is documented with a version history so that any study or reference can cite the exact version it was built on. Major revisions are recorded; the framework is never silently altered. As the framework matures, its version history will be published here so that the evolution of the model is fully traceable.