July, 2026.- Every year, Cannes celebrates the creativity you can see. The films, the stunts, the headlines that stop you mid-scroll. It’s a festival built on visibility and rightly so, because visible creativity still wins business, wins culture, wins awards.
Francisco Vigo CEO & Co-Founder, geoSurge and Cesar Christoforidis CEO & Founder Gravana
But sitting down with Francisco Vigo CEO & Co-Founder of geoSurge (a leader in the field of helping brands to be discoverable on LLMs) at The Culture Village in Cannes, a different kind of creativity came into focus. One that no consumer will ever consciously notice, but that may end up deciding which brands even get considered in the first place.
Call it invisible creativity. It doesn’t live in a campaign. It lives inside the model.
The new gatekeeper isn’t a search engine
For two decades, discoverability meant SEO. Keywords, backlinks, page authority a visible, gameable, well-understood system sitting on top of Google.
That system is being quietly displaced. As AI agents become the primary interface for shopping and research, product discovery is shifting away from search and browsing and toward agent recommendation. The brand doesn’t need to be found by a person scrolling a results page anymore. It needs to be remembered by a model.
This is where Francisco’s work gets interesting. geoSurge operates on what he calls Corpus Engineering™ the deliberate, technical work of interrogating and shaping how a brand sits inside an LLM’s parametric memory. Not a meta tag. Not a backlink. The actual internal representation a model draws on when it’s asked, in effect, “what should I recommend here?”
That’s not copywriting. It’s not even traditional SEO. It’s a form of creativity that never produces a visual, a slogan, or anything a brand’s customer will ever lay eyes on and yet it directly shapes whether that customer’s AI agent suggests the brand at all.
The market is already voting
geoSurge isn’t just a theory. The company raised $12m in a round led by AlbionVC, with participation from Play Ventures, Octopus Ventures, Celero Ventures, Boost Capital, Passion Capital and Tuesday Capital, a syndicate that reflects serious conviction in a company barely a year old.
AlbionVC’s investment manager Valérie Aelbrecht framed it plainly: geoSurge is building foundational technology for a category that will become increasingly important as AI systems shape more commercial decisions.
Francisco’s own framing of why that category matters is worth quoting directly: “A lot of the market is still thinking about AI visibility like SEO and citation tracking: we believe that’s fundamentally wrong. The real battleground is how models learn, understand, remember and represent brands over time.”
The timeline is shorter than most brands think
“The Discovery Journey Is Changing Rapidly”
Traditional Discovery vs. AI Agent Discovery, 2025–2030]
In 2025, traditional discovery still dominates, sitting close to 90%, with AI agent discovery a minor fraction. But the lines aren’t parallel – they’re converging fast. By 2027, the projection has traditional discovery at 72% and AI agent discovery already at 28%, and the trajectory keeps closing from there, with the two crossing before the decade is out.
That’s not a distant horizon. That’s roughly twelve months from the kind of inflection point where a brand invisible inside LLM memory becomes invisible to a meaningful share of its actual buying journey not because of a bad ad, but because nobody ever engineered it to be there.
Customers don’t search anymore. They ask.
That shift isn’t evenly distributed, and Latin America is one of the regions where it’s moving fastest.
“ChatGPT Will Hit 500M LATAM Users in 2026” User Growth, Jul 2023–May 2026
The growth curve is close to vertical. From a near-flat base in 2023, ChatGPT’s LATAM user base climbs past 100M by mid-2025 and is on pace to approach 500M by May 2026. That’s not a niche behavior shift that’s hundreds of millions of people across the region defaulting to a conversation instead of a search bar.
For brands in retail fashion, operating across the LATAM market, this isn’t an abstract trend to monitor. It’s a live audience that’s already asking an AI agent for recommendations, right now, in a market where “what should I buy” increasingly means “what does ChatGPT tell me to buy.” If a brand’s parametric memory hasn’t been engineered for that conversation, it’s not in the running not because of weak creative, but because it was never in the model’s frame of reference to begin with.
This is the second half of the urgency case: it’s not just that the crossover from search to AI agents is coming globally within roughly twelve months. In LATAM specifically, the audience has already arrived.
“They don’t send any traffic back”
In our conversation, Francisco put the economics of this shift in blunt terms. For two decades, the web ran on an implicit trade: publishers let Google crawl and index their content, and in exchange, Google sent people back to the website to convert, subscribe, or buy. That exchange, he argued, is breaking down.
“Google is crawling it, using it to train Gemini, and then they don’t send any traffic back,” he said. The content still gets consumed, it’s just consumed inside the model, with nothing flowing back to the source. As he put it, the question brands and creators now have to ask is blunt: if no traffic comes back, why keep creating content for the open web at all?
That question applies just as much to brand content as it does to publishers and creators. It’s the same logic driving the urgency behind the charts above; visibility on the open web doesn’t guarantee visibility inside the model a customer is actually asking.
Competing with generation itself
The same dynamic, Francisco argued, is already reshaping the creator economy and it’s a useful preview of what’s coming for brands. Platforms like Instagram and TikTok built their early value on a simple exchange: creators made the content, platforms distributed it, audiences supplied the attention, and creators got compensated through reach. That exchange depended on the platform needing creators to fill the feed.
It doesn’t anymore. As Francisco bluntly framed it, creators today aren’t only competing against other creators for attention they’re “competing with generation,” full stop. The platforms now have the means to generate hyper-engaging content themselves, without the creator in the loop at all.
It’s a sharp way of naming what’s at stake, and it’s the same fight brands are about to walk into. If platforms and models can generate the engaging surface themselves, the leverage shifts entirely to whoever controls the underlying memory and meaning the model is drawing from which is precisely the invisible layer geoSurge is built to engineer.
Why this should unsettle the creativity establishment
Cannes exists to reward work that’s seen. But the work now quietly determining whether a brand gets a seat at the table in the first place is unseen by design. It doesn’t sit in a portfolio. It doesn’t get a Lion.
That’s the disruptive part of this conversation: the industry’s definition of creativity hasn’t caught up with where the actual battle for attention has moved. Agencies and brand teams obsessing over the next visible campaign may be optimizing for a channel that’s shrinking, while a layer of strategic, technical, deeply creative work the kind that determines whether an LLM even considers you a credible answer goes almost entirely unrecognized.
The brands who understand this early who treat their presence inside AI parametric memory as seriously as they treat their visual identity are the ones positioning themselves for the crossover point already on the chart. The rest will find out about it the hard way, sometime around 2027 or 2028, when the traffic numbers stop being theoretical.
The real question for the industry
If creativity’s job has always been to be remembered, then the new creative frontier isn’t where humans are looking. It’s where the machines are listening.
The next great creative brief might not produce a single visible asset. And the next Cannes Lion-worthy idea might be one nobody outside a handful of engineers and strategists ever sees.
