IAB’s Gabilan: Agentic AI will Expose ‘Old Cracks’ in Ad Measurement
The humble impression may be living on borrowed time. When an AI agent browses the web, compares products and completes a purchase on a consumer’s behalf, what exactly did that consumer “see”? And if no human eye ever landed on an ad, does the impression even count?
These are the kinds of questions keeping measurement specialists awake at night as agentic AI reshapes how consumers discover, evaluate and buy products. The shift from human-driven browsing to agent-mediated commerce threatens to upend decades of established metrics, from clicks to viewability to last-touch attribution.
“Impression is not an impression anymore if it’s seen not by a human eye,” said Cintia Gabilan, svp, product development, IAB, in this video interview with Beet.TV at IAB CreatorFronts. “Human clicking will not be as it was because you just share a prompt and then you get something in return without needing to click in anything.”
The vocabulary problem
The advertising industry built its measurement infrastructure around a key assumption: humans see ads, humans click ads, humans buy things. Agentic AI disrupts every link in that chain. A consumer expresses intent through a prompt, an agent executes the rest, and the traditional touchpoints that marketers have relied upon for decades simply vanish.
Gabilan argues this creates an urgent need for new language and new standards. The industry lacks shared vocabulary for what success looks like when AI agents become the primary interface between brands and consumers. Without common definitions, each platform risks creating its own measurement silo.
“We need to rethink alongside all the legacy methodologies and how we can combine them,” Gabilan said. “Think MMM, incrementality, attribution. Not only how they harmonize with each other, but also on top of it, how the agentic web come along with all this non-language we don’t have, and the KPIs we need to discuss.”
CFOs want receipts
The pressure is coming from the top. Chief financial officers and chief marketing officers are increasingly demanding proof that their AI investments deliver tangible business results. Yet the industry has not even established baselines for evaluating AI tool performance, let alone the new advertising formats emerging inside large language models.
Emarketer forecasts that U.S. AI ad spending will more than double to $68.25 billion by 2030, with more than 80% of AI advertising in 2026 appearing next to AI content rather than inside conversational interfaces. That distinction matters: ads adjacent to AI Overviews behave differently than ads embedded in chatbot conversations, and measurement approaches will need to reflect those differences.
“There is an underrated phenomenon going in the industry where the CFOs and the CMOs are asking their teams, where is the ROI of all this AI tooling that we are paying,” Gabilan said. “There is not even a benchmark to analyze and evaluate all of that.”
Blurred lines everywhere
The measurement challenge arrives amid broader fragmentation across the advertising landscape. Gabilan pointed to creators becoming media businesses, video being reshaped by streaming and live commerce, and retail media integrating with other channels. Traditional category boundaries are dissolving, making budget allocation and performance comparison increasingly complex.
IAB’s own research released in September found that 72% of consumers now use general-purpose AI assistants for shopping decisions, with 56% preferring AI recommendations that incorporate creator perspectives. That convergence of creator influence, commerce and AI-mediated discovery represents exactly the kind of cross-channel complexity that existing measurement frameworks struggle to capture.
“There is no interoperability, there is no shared taxonomy, standards, consistency, and it’s still a very extremely fragmented landscape,” Gabilan said. “We are not short of work at IAB.”
Leading rather than following
Some industry voices have suggested it is too early to standardize around agentic advertising. Gabilan pushes back firmly on that view, arguing that waiting for the market to mature risks ceding control to individual platforms that will define measurement on their own terms.
IAB Tech Lab has already introduced AAMP 3.0, a set of protocols for standardizing how AI agents discover, negotiate and execute media opportunities. The organization has also published a framework for measuring visibility in the AI era, introducing concepts like the “4 P’s” of presence, prominence, portrayal and persuasion.
“I would rather have us leading ourselves into the future than being on the seat where we are just trying to adapt and fix the gaps,” Gabilan said. “It’s going to be a very critical year for us to come together and define that instead of letting it be defined by each LLM themselves without any shared goal and needs.”
The humble impression may be living on borrowed time. When an AI agent browses the web, compares products and completes a purchase on a consumer’s behalf, what exactly did that consumer “see”? And if no human eye ever landed on an ad, does the impression even count?
These are the kinds of questions keeping measurement specialists awake at night as agentic AI reshapes how consumers discover, evaluate and buy products. The shift from human-driven browsing to agent-mediated commerce threatens to upend decades of established metrics, from clicks to viewability to last-touch attribution.
“Impression is not an impression anymore if it’s seen not by a human eye,” said Cintia Gabilan, svp, product development, IAB, in this video interview with Beet.TV at IAB CreatorFronts. “Human clicking will not be as it was because you just share a prompt and then you get something in return without needing to click in anything.”
The vocabulary problem
The advertising industry built its measurement infrastructure around a key assumption: humans see ads, humans click ads, humans buy things. Agentic AI disrupts every link in that chain. A consumer expresses intent through a prompt, an agent executes the rest, and the traditional touchpoints that marketers have relied upon for decades simply vanish.
Gabilan argues this creates an urgent need for new language and new standards. The industry lacks shared vocabulary for what success looks like when AI agents become the primary interface between brands and consumers. Without common definitions, each platform risks creating its own measurement silo.
“We need to rethink alongside all the legacy methodologies and how we can combine them,” Gabilan said. “Think MMM, incrementality, attribution. Not only how they harmonize with each other, but also on top of it, how the agentic web come along with all this non-language we don’t have, and the KPIs we need to discuss.”
CFOs want receipts
The pressure is coming from the top. Chief financial officers and chief marketing officers are increasingly demanding proof that their AI investments deliver tangible business results. Yet the industry has not even established baselines for evaluating AI tool performance, let alone the new advertising formats emerging inside large language models.
Emarketer forecasts that U.S. AI ad spending will more than double to $68.25 billion by 2030, with more than 80% of AI advertising in 2026 appearing next to AI content rather than inside conversational interfaces. That distinction matters: ads adjacent to AI Overviews behave differently than ads embedded in chatbot conversations, and measurement approaches will need to reflect those differences.
“There is an underrated phenomenon going in the industry where the CFOs and the CMOs are asking their teams, where is the ROI of all this AI tooling that we are paying,” Gabilan said. “There is not even a benchmark to analyze and evaluate all of that.”
Blurred lines everywhere
The measurement challenge arrives amid broader fragmentation across the advertising landscape. Gabilan pointed to creators becoming media businesses, video being reshaped by streaming and live commerce, and retail media integrating with other channels. Traditional category boundaries are dissolving, making budget allocation and performance comparison increasingly complex.
IAB’s own research released in September found that 72% of consumers now use general-purpose AI assistants for shopping decisions, with 56% preferring AI recommendations that incorporate creator perspectives. That convergence of creator influence, commerce and AI-mediated discovery represents exactly the kind of cross-channel complexity that existing measurement frameworks struggle to capture.
“There is no interoperability, there is no shared taxonomy, standards, consistency, and it’s still a very extremely fragmented landscape,” Gabilan said. “We are not short of work at IAB.”
Leading rather than following
Some industry voices have suggested it is too early to standardize around agentic advertising. Gabilan pushes back firmly on that view, arguing that waiting for the market to mature risks ceding control to individual platforms that will define measurement on their own terms.
IAB Tech Lab has already introduced AAMP 3.0, a set of protocols for standardizing how AI agents discover, negotiate and execute media opportunities. The organization has also published a framework for measuring visibility in the AI era, introducing concepts like the “4 P’s” of presence, prominence, portrayal and persuasion.
“I would rather have us leading ourselves into the future than being on the seat where we are just trying to adapt and fix the gaps,” Gabilan said. “It’s going to be a very critical year for us to come together and define that instead of letting it be defined by each LLM themselves without any shared goal and needs.”
