IAB AI-Era Attribution Blueprint: conversions classed Human-Owned, AI-Assisted or AI-Executed

The IAB released the AI-Era Attribution Blueprint on 12 November 2026, a framework that classifies every conversion as Human-Owned, AI-Assisted or AI-Executed and uses that pathway to decide who may claim credit for the sale and what evidence the claim needs. The IAB announced the blueprint in a press release from New York and described it as working alongside existing attribution systems rather than replacing them. It is a framework and a call to action, not a finished measurement standard, and it comes from a US industry body.

What the IAB released on 12 November

The IAB said on 12 November 2026 that the blueprint is a framework for classifying AI's role in a conversion and for determining which parties can claim credit for it. It starts from three questions about each sale:

• Who or what created the intent?
• Who or what made the decision?
• Who or what completed the purchase?

The answers place the conversion in one of three pathways: Human-Owned, AI-Assisted or AI-Executed. The pathway then determines which parties are eligible for a "decision claim" or a "contributory claim", if any, and what evidence each claim requires. The IAB's wording is that a claim is not credit; it is the right to be evaluated for credit. That distinction matters. The blueprint does not hand out percentages of a sale. It sets the rules for who gets a seat at the table when credit is assigned and what they must bring.

The full document sits behind a free IAB account login. The public preview page lists the three pathway names, a framework for who can claim credit and what evidence is required, guidance on available evidence and current measurement gaps, and a section on transparency issues. The detailed pathway definitions are not in the public press release, so this article does not paraphrase them beyond what the IAB published openly.

Why the IAB built it

The IAB tied the release to its own buyer research. In the IAB 2026 Outlook Study September Update, 45% of buyers named comparing AI-driven and traditional customer journeys as their top measurement challenge, and 44% said they use branded search and direct traffic as proxies for AI influence they cannot track.

Caroline Giegerich, the IAB's VP of AI, said in the release that AI-influenced sales are sometimes reported as search or direct traffic, which can send the next marketing investment to the wrong channel. IAB CEO David Cohen said the blueprint is candid about what the industry cannot yet substantiate and invited the ecosystem to apply it.

The Similarweb numbers behind the problem

Digiday, which covered the release on 12 November 2026, reported Similarweb figures cited in the IAB report. Consumers who received a brand recommendation from an AI assistant were 2.5 times more likely to visit that brand's website within seven days. But only 8.8% of those visits came directly from AI referrals. About 55.9% were attributed to search and 19.9% to direct traffic.

That is the mechanism in plain numbers. The AI answer did the persuading, and a later search or a typed URL collected the credit in the analytics report. A budget review that reads only last-touch channel data would see branded search and direct performing well and could put more money there, which is the misallocation Giegerich described.

The table below collects the figures from the IAB release and Digiday's report.

What the IAB asks each party to do

The IAB's release set out a separate ask for each group in the chain.

• Brands should estimate the share of their conversions that fall in each pathway.
• AI platforms should disclose whether recommendations are paid or organic and what sources informed them, and label AI-originated traffic consistently.
• Agencies should apply one claim-and-evidence standard across platforms, separating platform-reported claims from independently verified ones.
• Publishers should identify the contributory claims their content creates and the evidence needed to support them.

The IAB said it will convene brands, agencies, AI platforms, retailers and publishers to share learnings as more data becomes available and to narrow the measurement gaps. It scheduled an "Attribution in the AI Era" webinar for 13 November 2026.

The publisher example

Digiday described how a contributory claim could work for a publisher. A publisher's review may shape a sale that an AI assistant recommends, which makes a contributory claim possible. But citations alone do not show how much the content influenced the answer, and use of content in training data is generally untraceable from outside the model. In other words, the blueprint gives publishers a vocabulary for the claim before it gives them the data to prove it.

Who built it and how long it took

According to Digiday, Giegerich said the work took about four months, and Digiday had first reported in August 2026 that the IAB was developing an AI attribution framework. Digiday named working group members including People Inc., Walmart, WPP Media, Emberos, Criteo, Bluefish, Microsoft, Brandlight, Tinuiti and Prudential Financial. Giegerich, speaking at a virtual IAB roundtable reported by Digiday, said the document became "more of a call to action to the industry."

The IAB said the blueprint extends its Measuring Visibility in the AI Era playbook, moving from how brands appear in AI platforms to how that presence connects to sales. Relevant Audience covered that playbook on 5 August 2026: the IAB released the 36-page framework on 3 August 2026 and said only 16% of brands systematically track AI visibility.

What the release did not say

Several things are missing, and they shape how much weight to put on the blueprint today.

• No tooling. The IAB did not release software, a tag, a reporting template or a GA4 configuration.
• No platform commitments. Neither the IAB nor Digiday said Google, OpenAI or Meta had agreed to the disclosure asks, although Microsoft was named in the working group.
• No standard status. Giegerich called it a call to action; the IAB did not present it as a certified measurement standard.
• No geography beyond the US body. The release does not mention Thailand or Southeast Asia.
• No public pathway definitions. The detailed definitions and evidence rules are in the gated document.

The wider measurement picture this week

Digiday placed the blueprint alongside other developments. The UK Competition and Markets Authority requires Google to give UK publishers clickthrough data for its AI search features by 3 December 2026. OpenAI is testing a dashboard showing how often businesses appear in ChatGPT answers and the clicks those appearances drive. Relevant Audience covered the OpenAI side on 7 November 2026, when the "Claim your business in ChatGPT" hub was reported, and on 11 November 2026, after OpenAI launched Businesses in ChatGPT in beta on 10 November with mention and click data. Digiday also cited a Profound analysis finding that one in three consumers had a relevant AI purchase conversation within a month that went unattributed, a dataset that excluded Google's AI Mode and AI Overviews.

Taken together, the pieces fit. Platforms are starting to release their own visibility data, regulators are pushing for clickthrough data, and the IAB is trying to set rules for how anyone should use that data to claim credit for a sale.

What to check in your own reporting

This section is analysis, not part of the IAB release. The blueprint's first ask of brands, estimating the share of conversions in each pathway, can start with data most advertisers already hold.

• Look at branded search and direct traffic trends next to your AI visibility work. If both rise while other channels hold flat, the Similarweb pattern suggests some of that growth may have started in an AI answer.
• Separate AI-assistant referrals in analytics. In GA4, a custom channel group for known AI assistant referrers makes the 8.8%-type slice visible instead of leaving it inside "Referral".
• Add a "how did you hear about us" question with an AI assistant option to lead forms and checkout. Self-reported data is imperfect, but it is the kind of independent evidence the blueprint asks for.
• When an AI platform reports conversions for its own ads, record it as platform-reported, separately from what your own analytics verifies, which is the split the IAB asks agencies to make.

What this means for Thai marketers

The blueprint is US-made and the IAB did not mention Thailand, so the following is analysis. The underlying problem does not depend on geography. Any Thai brand that people research in ChatGPT, Gemini or Google's AI features can see the same effect the Similarweb data shows: the AI answer persuades, and branded search or direct traffic gets the credit. Thai advertisers working on generative engine optimisation or testing ChatGPT Ads can borrow the three-pathway vocabulary now to describe results honestly, without waiting for a local standard. The practical step is to stop treating branded search growth as proof that branded search ads caused it.

FAQ

What is the IAB AI-Era Attribution Blueprint?

It is a framework the IAB released on 12 November 2026 that classifies each conversion as Human-Owned, AI-Assisted or AI-Executed and uses that pathway to decide who may claim credit and with what evidence. It works alongside existing attribution systems.

Is it an official measurement standard?

No. IAB VP of AI Caroline Giegerich said, as reported by Digiday, that it became "more of a call to action to the industry", and the IAB did not release tooling or certification with it.

What is the difference between a claim and credit?

A claim is the right to be evaluated for credit, not the credit itself, according to the IAB. The pathway decides whether a party can make a decision claim or a contributory claim and what evidence that claim needs.

Does it apply in Thailand?

The IAB did not mention Thailand; it is a US industry framework. The misattribution it describes, AI-influenced sales showing up as search or direct traffic, is not specific to any market, so Thai advertisers can apply the vocabulary voluntarily.

Do I need to change anything in GA4?

Nothing is required, because the blueprint comes with no GA4 changes or platform mandates. Separating AI-assistant referrals into their own channel group is a sensible step if you want to estimate your pathway mix.

If your branded search and direct numbers are growing and you are not sure what is driving them, Relevant Audience's GA4 team can help set up reporting that separates AI-assisted traffic from the channels that usually take the credit.