Why we invested in geoSurge

By Valerie Aelbrecht, Investment Manager

AI / Data

Posted
July 3 2026

Search built the internet economy. For more than two decades, appearing on page one of Google was among the most valuable pieces of digital real estate a brand could own. Today, something structurally important is changing in how people find information, and how companies get discovered.

Consumers are increasingly turning to AI systems such as ChatGPT, Gemini, Claude and Google’s AI Overviews to research products, evaluate brands and make decisions. ChatGPT for example has now crossed the 1 billion monthly-user milestone as of May 2026. According to McKinsey, AI-powered search has already become a primary source of information for many consumers, with hundreds of billions of dollars of purchasing decisions expected to be influenced through AI search experiences over the coming years.

This matters for every company that depends on being discovered, understood, and chosen online. The mechanisms that shaped visibility in the Google era like links, rankings, structured content, are no longer sufficient on their own. Increasingly, users are not browsing lists of links; they are receiving synthesized answers. A new visibility economy is emerging, and most brands have little understanding of how they appear within it. This is the shift geoSurge has been built for, giving brands visibility into, and helping improve , how they are represented in AI-generated answers.

The problem: brands are disappearing from AI memory

For the first time, brands risk becoming materially underrepresented in AI-generated recommendations. In the Google era, companies worried about dropping from position 1 to position 10. In the AI era, they risk not appearing at all.

If a model does not associate a company with a category, product or problem, it may simply never mention it. And unlike traditional search, there is often no obvious explanation as to why.

This represents a structural shift in how discovery works. Marketers today have virtually no visibility into how LLMs perceive their brand. The frameworks, tools and attribution approaches built for SEO do not translate cleanly to AI systems. Unlike Google’s deterministic rankings, LLMs are probabilistic: a brand’s presence in responses varies across models, prompts, geographies, and model versions. And with each new model release cycle, the slate is wiped.

The solution: corpus engineering, not elevated SEO

Most companies operating in the GEO space focus on measuring outputs: tracking citations, monitoring prompts and helping brands understand how often they appear in AI-generated responses. That is useful, but we believe much of that functionality will become increasingly commoditised over time. It is not difficult to imagine leading model providers eventually offering their own visibility and analytics tooling, much as Google created Search Console for the web era.

There is a deeper problem with the current category, however. Many tools effectively mark their own homework e.g. optimising the very content they create, tweaking prompts to surface more favourable outputs, and measuring performance against signals they themselves influence. That is not a neutral measure of brand visibility but more of a feedback loop. Similarly, approaches that treat GEO as elevated SEO, e.g. producing more content, adding more structured markup, chasing retrieval rankings, are applying a deterministic playbook to a fundamentally probabilistic problem. The category needs a different starting point.

Most of the category today is focused on observing what models say. geoSurge is focused on understanding why models say it.

The company starts from a different premise. There are two forms of AI visibility. The first is dynamic visibility: what models retrieve from the web in real time. The second is parametric visibility: what models have already learned through training. geoSurge knows the long-term strategic challenge lies in understanding and improving how brands are represented within the models themselves. As frontier AI systems mature, learned representations will shape how answers are generated more, and more. Understanding those representations, and the signals that impact them, requires a fundamentally different approach from traditional SEO or content optimisation.

geoSurge operates at the model layer as opposed to the content layer, with a proprietary approach called Corpus Engineering™. It is designed around understanding and improving the signals that shape how models learn and represent brands over time, not just how they appear in retrieval. We believe that distinction matters, as this requires a different technical stack, a different data strategy, and a different view of where long-term value will accrue. And so, apparently, do customers. We have seen brands move directly from visibility and content tools to geoSurge the moment they saw the product in action. If you are thinking about your own AI visibility strategy, it is worth seeing for yourself: geosurge.ai

The team: a technical tandem

The best technical founders often see the world through a different lens. Francisco and Jons fit that profile.

What impressed us most was not just their technical depth, but how they frame the problem itself. Where much of the market approaches AI visibility as an extension of SEO, Francisco and Jons approach it as a question of model behaviour and architecture. That perspective leads to fundamentally different product decisions, a different research agenda and, in our view, a more defensible position over time.

Francisco brings the rare combination of data scientist rigour and commercial instinct – someone who understands model behaviour at a probabilistic level and can translate that into enterprise value. Jons has been building neural networks since his teens and brings systems-level thinking that is rare even within technical founding teams. In a market moving fast and attracting a lot of noise, they have demonstrated a willingness to challenge prevailing assumptions about the category and pursue a genuinely differentiated approach.

Why we invested – and why now

AI-powered discovery is moving from experimentation to mainstream behaviour. According to McKinsey, roughly half of consumers already use AI-powered search, and brands that fail to adapt risk losing a meaningful share of traditional search visibility over time.

Unlike SEO, which largely operated against a single dominant platform with relatively transparent ranking mechanisms, GEO operates across multiple models with fundamentally different behaviours. The technical complexity is significantly higher, creating a category where deep technical advantages have the potential to compound over time.

We believe every organisation will eventually need to understand how it is represented inside AI systems, just as every organisation today understands how it performs in search. geoSurge is building the infrastructure layer for that future. We are proud to lead the company’s $12 million seed round alongside Boost VC, Celero Ventures, Passion Capital, Octopus Ventures, Tuesday VC and an exceptional group of angels, including executives from Google DeepMind, Microsoft AI and Signal AI.
Valerie Aelbrecht, Investment Manager

Sources

ChatGPT app hits 1 billion monthly active users in record time, data shows | Reuters

Winning in the age of AI search | McKinsey

News