Time-to-citation curves

“When will we see results?” has no single answer, because AI visibility moves on two clocks — the retrieval layer (hours to weeks) and the model layer (model-release cycles). This project instruments releases so you can answer the question with your own numbers instead of vendor averages, and quietly upgrade every expectation-setting conversation.

Why this, mechanically. Published timelines are directional and mostly come from vendors selling tracking tools; internal baselines matter more. Treating every significant release as a measured event with a first-citation clock per surface builds timing curves specific to your catalog and markets — the only real scoreboard for how long the work actually takes.

Deliverable. Instrumentation on the next 20 significant content, schema, or feed releases across the archetypes; days-to-first-citation recorded per surface and top market; and the org’s own published timing curves after a quarter.

Finish line. The timing curves exist and are attached to the sequencing guide and to expectation-setting conversations with leadership.

This is the slow-compounding end of measurement, and it depends on the prompt corpus baseline being live to detect first citations.