Most tools that measure brand presence in AI do the same thing: they count how many times ChatGPT or Gemini mentioned your brand in a response. They call it visibility. Some add sentiment. A few more ambitious players add position in the response.
These are all presence metrics. Not attention metrics.
The difference is fundamental and has a direct impact on the budget decisions you make every quarter.
Presence is not attention
When a user asks an AI model a question, they receive a response ranging from a few hundred to several thousand characters. Your brand might appear in the first sentence, somewhere in the middle of a paragraph, or in a list at the end. Tools measuring visibility treat all three situations the same way: the brand appeared, count it.
Behavioral research in generative environments shows something different. Attention in AI chat follows its own decay curve. Information in the first sentence of a response attracts many times more attention than the same information in the fourth paragraph. A brand mentioned with context that justifies a choice works differently than a brand listed among options to consider. Sentiment is not binary: a positive context with low cognitive load produces a different result than a positive context in a response the user abandons after three sentences.
None of the dominant tools measure this. They sell you a mention counter with an "AI analytics" label.
Why AI requires a separate measurement standard
AI chat is a new communication environment with no equivalent in media history. It is not a website: there is no scroll, no headers, no visual structure guiding attention. It is not video: there is no autoplay, no sound, no editing rhythm. It is not display advertising: there is no time-based exposure, no viewability in the classical sense.
An AI response is a linear stream of text generated in real time, read by a user who asked a very specific question and expects a very specific answer. The level of intent here is the highest of any digital channel. The tolerance for content that does not match that intent is the lowest.
This means attention mechanisms work differently here. The decay curve is steep and fast. Position matters disproportionately more than in any other medium. The quality of the justification determines whether a brand stays in the user's memory or disappears with a closed browser tab.
The methodological conclusion is direct: AI chat must be measured using a separate standard. That standard should define not only what to measure, but how to weight position, context, attention quality, and decay over time.
What the IAB standard changes
This is precisely what is now being addressed in IAB standardization work. The chapter dedicated to AI environments within the attention measurement standard carries normative status: it applies to all market participants who want to operate within a recognized industry framework.
This is a structural change, not a technical one. It means that in a year or two, the market will not divide between those who measure AI presence and those who do not measure at all. It will divide between those who measure according to the standard and those whose data is not comparable to anything.
Media agencies advising clients on tools that measure only visibility will need to redesign their reports. Publishers monetizing AI citations will need a metric they can negotiate at the table with a buyer. Brands optimizing AI presence today without measuring attention are optimizing for the wrong goal.
BAX Index and the AI environment
BAX measures attention in AI responses across four dimensions: visibility, position, sentiment, and exposure quality. Each dimension is weighted according to the attention decay curve specific to generative environments, calibrated on data from more than 300 AI engines.
The BAX Index is not a mention counter. It is a measure of the real attention value your brand earns in the channel that is becoming the first point of contact for millions of purchasing decisions.
The AI measurement market is selling you presence. We measure attention. These are not the same thing.