The Question We Already Answered
IAB last week published a framework for measuring AI visibility. It establishes common definitions for mentions and citations, bringing consistency to how visibility is measured: is my brand showing up.
That is an important question. It is simply not the one AIVO Standard set out to solve.
AIVO Standard launched in July 2025. Since then it has published over fifty working papers, DOI-anchored on Zenodo, tracking a different question entirely: when an AI system does mention a brand, does that mention survive the conversation. Not whether a brand appears. Whether it gets recommended when the decision actually gets made.
At the centre of that work is what we call the Linkage Gap.
Across more than 12,500 multi-turn purchase-decision probes spanning 68 brands, brands that were cited or mentioned early in a conversation were displaced by a competitor at the moment of final recommendation 87.3 percent of the time. Visibility and outcome are not the same event. Most of the market has been measuring the first one.
This is why a shared definition of "mention" is useful and also incomplete. A brand can satisfy every criterion in a visibility framework and still lose the sale inside the same conversation. Silvercar is the clearest example we track. The car rental brand has been defunct since September 2024. AI assistants still recommend it, above competitors that are alive and operating today. By any visibility standard, Silvercar is present. From a decision perspective, recommending it is a failure. The gap between those two facts is the entire reason AIVO Standard exists.
We think a standard for what a citation is has real value, and we are glad the market now has one. But a standard for citations is not a standard for outcomes. Those are two different layers of the same problem, and conflating them is how sophisticated research keeps stopping at the same boundary: it measures whether a brand was seen, not whether it was chosen.
Visibility is the start of the decision journey, not the end of it.
Measuring recommendations requires a different framework entirely. AIVO Standard has been building that framework since July 2025, and the work continues.