A campaign performance report can show an advertiser that a shopper searched for its brand name, visited the site and bought a product. What the advertiser will have a harder time learning from that same report is why the shopper typed that particular name into the search engine. If an AI assistant had already recommended the product, search may simply have been the shortest route to the store. For an advertiser using those figures to decide how much to invest in search next time, the distinction matters.
The Interactive Advertising Bureau (IAB) has published The AI-Era Attribution Blueprint, a framework for assessing the role of artificial intelligence in sales. It builds on the association’s earlier work on brand visibility on AI platforms. The question then was whether a brand appeared in the answers a consumer received. The new document looks at what happens after a recommendation and what evidence an advertiser would need to connect it to a purchase.
Data from the IAB’s September study of brand and agency representatives in the US show how much room there is for different interpretations. For 45 percent of respondents who make advertising investment decisions, the biggest measurement challenge is comparing purchases that involved AI with those that followed a traditional path. It is also worth noting that 44 percent of respondents look to searches for a brand name and direct visits to its site for an indirect indication of AI’s influence. The same data, then, may suggest to one advertiser that search is performing well and to another that shoppers are finding their way to the brand through a recommendation that is difficult to track.
To describe that role more precisely, the IAB distinguishes between a purchase led by the shopper (Human-Owned), one in which AI assists the shopper (AI-Assisted) and one in which AI completes the transaction (AI-Executed). The distinction between the second and third is particularly important. An assistant that helped a shopper choose a product played a different role from one the shopper entrusted with making the purchase. And knowing who completed the transaction still does not tell us whose recommendation ultimately determined the choice.
Publishers also have a place in this framework. A product review can serve as the source for an AI assistant’s recommendation, even when the shopper takes another route to the store. Under the new framework, the publisher then has grounds to ask for the contribution of its content to be assessed. How well it can substantiate that claim depends, among other things, on whether the platform discloses which sources it used for its answer. Much of the data needed to verify such a claim is currently available only to the platforms or is never recorded at all.
The IAB is therefore asking AI platforms to disclose the source of a recommendation and whether it was paid or organic, and to label visits originating from their platforms consistently. It suggests that agencies assess claims from different platforms against the same criteria. For a brand that entrusts its budget to an agency, this would also mean a concrete change in reporting: alongside the sales figures attributed to a given channel, the brand could see whether that channel’s contribution had been independently verified or reported by the platform where the money was spent.
