AI in MarketingAI for SEO & AEO
IAB Publishes First Guidance on Measuring AI Search Visibility
IAB's new guidance organizes AI search visibility into four measurable layers and sets a query-volume threshold for data reliable enough to guide budget decisions.
Key takeaways
- IAB published 'Measuring Visibility in the AI Era' on Monday, introducing four metrics: presence, prominence, portrayal and persuasion.
- IAB deliberately avoided calling it a 'standard' or 'framework,' saying AI search behavior is too unstable for either label yet.
- Anything under 50 queries counts as merely 'exploratory' data and cannot characterize a category, per the guidance.
- Only 'decision-grade' measurement, tested across larger query volumes and more platforms, should inform budget allocation.
- Publishers should watch 'persuasion' metrics closely since AI Overviews and similar tools are already cutting into referral traffic.
Four metrics, no standard yet
IAB published guidance Monday on how brands and publishers should track their visibility in AI search results, the first cross-industry attempt to bring order to a measurement landscape that currently has none. The document, "Measuring Visibility in the AI Era," organizes tracking around what IAB calls the 4 P's: presence, prominence, portrayal and persuasion, according to reporting from AdExchanger.
IAB's VP of AI, Caroline Giegerich, told AdExchanger the guidance is neither a "standard" nor, despite an earlier plan, a "framework." A standard needs stability, and marketers are currently in what she called a "mass transition space." The name change to avoid "framework" was, in her words, so the industry doesn't end up with 100 of them.
- Presence: how often a brand appears in AI search queries, or how often a publisher's content gets cited.
- Prominence: where a brand or citation lands in results, and whether it's highlighted or buried in a list of similar mentions.
- Portrayal: sentiment and accuracy, split between declining hallucination rates and factual inaccuracies from outdated data sets, which Giegerich said is the thing that would keep her up at night.
- Persuasion: whether a citation actually drives a click, measured through post-citation click-through rate.
The 50-query floor
The guidance also splits measurement into 'directional' and 'decision-grade' tiers. Directional data, like manually searching a brand name a few dozen times and eyeballing results, is treated as a rough signal at best. Decision-grade data requires rigorous testing across a larger query volume and multiple platforms, and IAB says only that tier should inform budget decisions.
Volume IAB classifies as merely 'exploratory,' unable to meaningfully characterize a category
IAB via AdExchanger, 2026
Why it matters now
Persuasion is the metric to watch if you run a publisher, since AI search tools including Google's AI Overviews are already answering queries without sending users anywhere. Marketers building out AEO programs should pair this guidance with our AEO playbook for getting cited by AI search, and cross-check it against the gap between AI brand recognition and actual citation rates, since knowing a brand and citing it are clearly not the same thing to a model.
Factual inaccuracies would be the thing that would keep me up at night.Caroline Giegerich, VP of AI, IAB
Track how AI search is reshaping your brand's visibility and budget decisions.
Advertiser disclosure: some links in our articles are affiliate links, and CMO Mag may earn a commission or referral fee if you sign up or buy through them, at no cost to you. It never affects our editorial coverage. See our advertising & affiliate policy.
More in AI in Marketing
View all
OpenAI Sets Ad Credit Rules as ChatGPT Ad Load Doubles
Six months into its ad business, OpenAI has quietly published the fine print on how brands can apply ad credits, even as its ad load per user more than doubles.
The AEO Playbook: How to Get Cited by AI Search in 2026
AI models recognize 96% of brands but cite almost none of them. Here's the mechanical reason why, and the content and structured-data playbook that actually moves citation rates.
AI Knows 96% of Brands, Cites Almost None of Them
A Q2 2026 study of eight AI platforms finds a wide gap between brand recognition and brand recommendation, and it's a warning for anyone counting on AI search for demand generation.
Discussion
No comments yet. Be the first to say something worth reading.