AI is not arriving at the Mexican digital ecosystem: it's already inside
On March 26, we presented our annual Digital Trends report from Comscore. As every year, the task of gathering data and translating it into narrative forces us to pause and ask ourselves what is truly new and what is continuity. This year, the answer was clear: artificial intelligence has stopped being an emerging trend and has become infrastructure. And that changes the questions we need to ask ourselves as an industry.
The internet user's journey is no longer what we knew
One of the findings that made me reflect the most relates to something seemingly simple: how we search. Or, better said, how we express an intention. For years, the process was predictable: a person had a need, wrote a query in a search engine, browsed results, and eventually made a decision. Today, that funnel is being reconfigured by artificial intelligence assistants.
Comscore's data illustrates this precisely. Access to the top ten global retailers —including Amazon, MercadoLibre, and Walmart— through ChatGPT from desktop devices grew by 181% between September 2024 and September 2025. This is not a minor detail: it speaks of a consumer who no longer clearly distinguishes between searching for information and making a purchasing decision. Intention and action are compressed into a single conversation with an AI tool.
This opens up a huge strategic question for brands, agencies, and media in Mexico: if the entry point to content and products is shifting towards AI assistants, what does it mean to appear —or not appear— in that space? SEO was for years the answer to that question in the traditional search environment. Today, the conversation is migrating towards what we call GEO: optimization for generative environments. And what Comscore is beginning to measure in that dimension —which domains AI assistants cite, how often, in what categories— is precisely the compass that many still do not know they need.
More users, more platforms, more complexity
In Mexico, 33% of internet users already use AI assistants from desktop devices, with a growth of 6 percentage points in just one year. That number coexists with a digital audience of 78 million unique visitors and a total consumption of 70.4 billion hours on the internet during 2025. Audiences are not shrinking; they are being redistributed and diversifying in their habits.
Streaming on Connected TV continues to gain ground. The time spent consuming YouTube on CTV in Mexico grew by 26% between January 2024 and January 2026. And 60% of CTV adoption in Latin America confirms that the big screen is no longer synonymous with linear television. For advertisers, this means that the same audience may be present at five different contact points before making a decision, and understanding that cross-platform journey is not an analytical luxury: it is an operational necessity.
Artificial intelligence, in this context, not only changes consumer behavior. It also begins to provide the tools to understand it better. The AI Rankers, for example, allow analyzing how visible a brand is within the responses generated by the assistants, building a metric of incremental potential audience that very recently did not exist. Measuring AI is no longer an academic exercise; it is part of market intelligence toolkit.
The World Cup as a moment of truth
The World Cup kicks off on June 11 in the United States, Canada, and Mexico, the most-watched sporting event on the planet. And, for the first time, a substantial part of its audience will experience it in a deeply digital way: through social media, streaming platforms, voice assistants, and yes, conversations with AI tools that will help find where to watch the match, what to eat during the game, or what jersey to buy.
The World Cup is, in this sense, the real moment for everything we have discussed this year. Brands that understand the new fan journey —fragmented, cross-platform, mediated by algorithms and assistants— will have a real advantage. Those that operate with the previous cycle's measurement and planning models will likely feel that something doesn't add up, even though the reach numbers may seem correct.
For Comscore, the World Cup also represents an opportunity to understand something that quantitative data alone doesn't always capture: the emotional and behavioral complexity of the fan. That's why we are developing, together with IAB Mexico, a specific study on the World Cup fan —combining surveys, expert interviews from the ecosystem, and digital behavior analysis— that seeks to precisely answer where the fan is, through which channel they experience it, what moves them to interact with a brand, and how that engagement translates into concrete decisions. Not to sell space, but to better understand a cultural phenomenon that, in Mexico, has dimensions that no global model fully captures.
An opportunity for those who measure well
I would conclude with something that I think it's important to point out from the position of someone who provides data to the industry: artificial intelligence is not a threat for those who measure well. It is, in fact, an opportunity. Because as consumer behavior becomes more complex —more platforms, more touchpoints, more algorithmic mediation— the need for independent, rigorous, and cross-platform measurement becomes more critical.
Mexico has a mature digital ecosystem, with large audiences, a very high affinity for categories like sports, retail, and entertainment, and a World Cup at home. The conditions are there. The question is whether as an industry we will make the decisions that this moment demands based on the data.