Where the IAB’s AI visibility framework ends — and the next challenge begins

The IAB’s new “Measuring Visibility in the AI Era” framework arrives at an important moment. More than 20 companies now offer AI visibility measurement, per the report. They often use different methodologies, prompt libraries and scoring systems, meaning that two providers can produce materially different answers about the same brand.

As my colleague Andrew Bolton wrote in a recent editorial in ADWEEK, “I have one client who uses four different LLM visibility tools simultaneously. Each one tells her something materially different about where her brand stands.”

The IAB’s response to this increasingly common situation is to establish common ground: a shared vocabulary, standards for measurement quality, disclosure requirements and practices for dealing with the inherent instability of AI-generated results. It’s similar to the trade organization’s move a dozen years ago to establish standards for ad viewability, which at the time were similarly muddled by different companies’ varying definitions and terminology.

That is an important step forward. But when you view it alongside Knotch’s research into AI-influenced customer journeys, it also highlights a bigger measurement challenge: Knowing whether you appear in AI is not the same as knowing what that visibility is worth.

1. IAB Takeaway: The industry needs a common language for AI visibility

The centerpiece of the IAB framework is its “4 P’s”: Presence, Prominence, Portrayal and Persuasion. Together they move from whether a brand appears, to where it appears, how it is characterized, and ultimately whether that visibility drives action.

That standardization is overdue. “Mention rate,” “citation rate” and “share of voice” are only useful if buyers understand how they are being calculated and can compare results across providers.

Knotch POV: Visibility is the beginning of measurement, not the end

The risk here is that standardized visibility could become yet another vanity metric. (Again, there are echoes of what happened with viewability standards here.)

Knotch’s research suggests that what can be directly observed through citations dramatically understates AI’s actual influence. While only around 0.13% of site traffic comes directly from LLM citations, 30–40% of visitors say they used an AI tool before arriving at the site.

Those visitors typically appear in conventional analytics as direct or organic traffic. The LLM may have shaped their consideration, intent and eventual decision without ever appearing in the referral path.

So Presence and Prominence matter. But the more consequential question is: What happens because of that presence?

2. IAB Takeaway: Not all AI visibility data should be trusted equally

One of the strongest parts of the IAB framework is its distinction between Directional and Decision-Grade measurement.

Directional data can identify trends and emerging signals. Decision-grade data requires greater rigor – around elements like sample size, query volume, prompt coverage, testing cadence, reproducibility, validation and platform coverage – before it should inform budget or strategy.

The IAB also makes an important point about false precision. AI responses are probabilistic: the same query can generate different answers, citations and brand recommendations. A reported “22% share of voice” therefore should not be treated like a deterministic search metric.

Knotch POV: Make the data decision-grade by connecting it to first-party outcomes

This is where Knotch would take the IAB’s principle one step further.

The IAB itself says directional data becomes more useful for higher-stakes decisions when signals converge across sources and are corroborated by first-party analytics.

That connection is a crucial one. Knotch has found that AI-influenced visitors are four times more likely to convert, while directly referred LLM journeys can behave very differently from traditional traffic.

A movement in citation rate or share of voice, therefore, becomes far more meaningful when it can be connected to what people subsequently actually do: how they engage, which content they consume, whether they progress through the journey and ultimately whether they convert. Visibility tells you whether you're being found, but first-party data tells you whether being found ended up mattering.

3. IAB Takeaway: Persuasion is where visibility becomes business value

Importantly, the IAB recognizes this problem. Its fourth P, Persuasion, asks the right question: “Does visibility drive action?”

Unfortunately, this is also where their framework is least developed – at least for now. The IAB defines Recommendation Strength and Post-Citation CTR, while explicitly positioning Persuasion as a bridge to a forthcoming attribution framework. It acknowledges that Post-Citation CTR depends on platform-level data that isn't consistently available. This should improve in the future, but we don’t have a timeline for when it will happen.

Knotch POV: Don't wait for perfect attribution

If AI's influence were limited to identifiable clicks from ChatGPT, Gemini or Perplexity, platform-side CTR might eventually provide much of the answer. But our research suggests the opposite: the vast majority of AI influence is invisible to conventional referral analytics. That means marketers need another way to expose it.

Knotch has been testing a first-party approach with our enterprise clients. We ask visitors via our Knotch Cards whether they used AI before coming to a site, then connect that declared behavior to their subsequent journeys, engagement and conversion.

The result is not a perfect attribution. But waiting for perfect attribution risks ignoring this rapidly growing influence on customer decision-making.

As a result, the emerging measurement model needs to connect GEO visibility,  SEO, first-party audience signals, and  site behavior with business outcomes – rather than relying on any single source to explain the journey.

4. IAB Takeaway: Transparency matters more than a single authoritative score

The IAB’s framework is particularly strong on provider transparency. It argues that buyers should understand how prompt libraries are constructed, where queries come from, which platforms are measured, how frequently tests are run, and how results are validated. It even warns that a provider with excellent reproducibility but a poorly constructed prompt library can produce “precise answers to the wrong questions.” It also recommends reporting platforms separately rather than allowing an aggregate score to conceal materially different performance across different LLMs.

Knotch POV: The answer isn't to choose one measurement system, but to connect them all

The AI customer journey now crosses systems that were never designed to talk to one another. GEO tools measure AI visibility, SEO tools measure search behavior, web analytics measure onsite activity, first-party audience data can reveal intent and influence, and conversion systems measure outcomes – but no one tells the whole story.

Our take at Knotch is that there is no definitive AI metric, and we shouldn’t be trying to anoint one. Rather, we should be connecting what’s out there so that marketers can understand the relationship between what AI says, what audiences see, what they do next, and what ultimately creates business value.

That is fundamentally the Content Intelligence problem that Knotch has been working on since long before AI and LLMs became concerns for marketers. Our mission has always been to connect  content performance across the journey instead of optimizing around isolated surface metrics. With our Knotch One product, for example, we combine engagement, audience sentiment, and journey signals to connect individual content experiences to downstream conversion. More recently, we launched Ace to extend that mission toward serving structured, semantically rich, citable content to AI agents themselves.

From measuring AI visibility to improving it

The IAB deliberately draws one final boundary: Its framework covers measurement, not GEO or AEO optimization. That makes sense for a standards body. But for marketers, measurement ultimately has to lead to action.

Our work suggests that content increasingly needs to work for two audiences: humans making decisions, and machines helping them make those decisions. Our research has found that LLMs favor clearly structured, authoritative passages that directly answer real audience questions. That’s potentially a way to rethink your entire website.

And it means that the next question after “Are we visible?” has to be “What should we change?”

The IAB has given the industry a much-needed common language for measuring AI visibility. The next challenge is connecting that visibility to the customer journey and ultimately to business outcomes. Because in the AI era, being visible is just the beginning of the journey.

‍

Published on September 2, 2026