The Interactive Advertising Bureau published a 36-page measurement framework called "Measuring Visibility in the AI Era" on 3 August 2026, and the document states that only 16% of brands systematically track AI visibility today. PPC Land reported the release on 4 August 2026.
The framework does not rank tools and does not tell anyone what to buy. It names the problem underneath the market: the industry has no agreed definition of the things it is trying to count.
What the IAB published on 3 August 2026
The IAB released "Measuring Visibility in the AI Era" on 3 August 2026 as part of Project Eidos, its measurement modernisation programme, and PPC Land wrote it up on 4 August 2026. The document is dated August 2026 and sets out how to measure the way brands appear in AI-generated answers. It is not a product, a certification or a vendor league table.
Its diagnosis is definitional rather than technical. According to the framework there is no agreed definition of a mention, no standard for what counts as a citation, and no shared method for judging whether a tool's output is rigorous enough to inform strategy. If two teams cannot agree on what a citation is, their numbers cannot be compared, whatever software produced them.
Only 16% of brands track AI visibility, and the figure is the IAB's own
PPC Land quotes the document directly on 4 August 2026: "According to the document, only 16% of brands systematically track AI visibility today, and the absence of a common standard is part of the reason." The IAB is offering a cause alongside the number, arguing that brands are not measuring partly because there is nothing stable to measure against.
The source does not state the sample size, the methodology or the geographic scope behind the 16%. It is the IAB's own figure, reported inside the IAB's own document, and neither the trade body nor PPC Land published the underlying research. Read it as directional, not as a benchmark.
Why two vendors can measure the same brand and disagree
The IAB framework counts more than 20 companies selling competing AI-visibility tools, and says they operate, in the document's words, "using different query sets, different platform coverage, different scoring rubrics and different definitions". Two vendors measuring the same brand, in the same category, in the same week can therefore return materially different results while each stays internally consistent. A share-of-voice score has no meaning outside the query set that produced it, and most query sets are the vendor's own.
Directional measurement and decision-grade measurement
The most portable part of the 3 August 2026 document is a two-tier split. The IAB describes directional measurement as useful for monitoring trajectory but insufficient for budget decisions, and decision-grade measurement as meeting higher standards for sample size, reproducibility and documentation. Those labels are worth adopting internally before any tool comparison starts, because a lot of what gets demoed is directional data being read as decision-grade evidence. The table below sets out what the framework contains and what it leaves alone, drawn from the IAB document as reported by PPC Land on 4 August 2026.
| Element of the framework | What the IAB document states |
|---|---|
| Brands systematically tracking AI visibility | 16%, the IAB's own figure, published 3 August 2026, no sample size or methodology disclosed |
| Document and programme | 36 pages, dated August 2026, part of Project Eidos |
| Directional measurement | Good for monitoring trajectory, insufficient for budget decisions |
| Decision-grade measurement | Meets higher standards for sample size, reproducibility and documentation |
| Vendor market | More than 20 companies sell competing tools using different query sets, platform coverage, scoring rubrics and definitions |
| Scope | Organic, non-paid AI visibility only; paid placement measurement named "an urgent adjacent priority" and not addressed |
| What it will not do | Does not rate providers, prescribe tools or build measurement systems |
What the framework refuses to do, and why that is not an accident
The IAB states that the framework covers "organic, non-paid AI visibility measurement only" and names paid placement measurement "an urgent adjacent priority" that this document does not address. It also declines to rate providers, prescribe tools or build measurement systems.
That restraint has a structural reason worth stating plainly. The IAB is a trade body funded by ad-industry members, and the working group behind this framework includes measurement providers. A document written partly by vendors was never going to rank vendors, so the supplier evaluation stays with the buyer.
What this means for Thai marketers
The source does not address Thailand and claims nothing about Thai data. For Thai advertisers the practical read is that AI-visibility dashboards are being pitched into this market with no shared basis for comparing them, and the framework's real contribution is a disclosure checklist to put in front of a vendor before signing anything.
Ask which prompts were run, on which assistants, how often, how a mention is defined, how a citation is defined, and whether the same query set can be rerun next quarter for a comparable number. A supplier that cannot answer is selling monitoring, which is fine to buy as long as it is budgeted as monitoring. Reporting on generative engine optimisation and AI SEO is more useful when each number carries the tier of evidence behind it.
Frequently asked questions
Is the 16% figure based on research I can check?
The source does not state the sample size, methodology or geographic scope behind the 16%. It is the IAB's own number, reported inside its own document, so it works as a directional signal rather than as evidence in a business case.
Does the framework tell me which AI-visibility tool to buy?
No. It deliberately does not rate providers, prescribe tools or build measurement systems. It gives a vocabulary and a set of quality criteria, and leaves the vendor choice with the buyer.
Does it cover Thailand or Thai-language AI results?
The source does not state whether any Thai market data is included, and nothing in the document is presented as Thailand-specific. Any application to the Thai market is a reader's inference, not an IAB claim.
Does it cover paid placements inside AI answers?
No. The IAB states the framework covers "organic, non-paid AI visibility measurement only", and calls paid placement measurement "an urgent adjacent priority" it does not handle. The source does not say when a paid-side framework might follow.
Do I have to do anything now?
Nothing here is mandatory. This is a voluntary framework from a trade body, not a platform rule, and no deadline attaches to it. The low-cost step is to adopt the tier language internally and ask any current supplier to document its query set and definitions.
The useful part is the vocabulary
The IAB has not solved AI-visibility measurement and does not claim to have. What the 3 August 2026 framework supplies is a shared way to talk about how much a number is worth, which is the missing piece when more than 20 vendors report different figures for the same brand. To have an existing AI-visibility setup reviewed against that standard, the SEO team at Relevant Audience can walk through it with you.
