geoSurge: $12M Seed for Corpus Engineering in AI Search

geoSurge is a London-based AI visibility company, founded in 2025, that helps brands and agencies measure how AI models describe them. It raised a $12 million seed round led by AlbionVC on July 3, 2026. Its trademarked method, Corpus Engineering, targets what models learn during training as well as what they retrieve live. Francisco Vigo is chief executive.

Last updated: November 1, 2026

What are the key facts about geoSurge?

geoSurge is a seed-stage company that sells AI visibility monitoring and a method it calls Corpus Engineering. The table lists its core facts from the July 2026 funding announcement and coverage of it.

How much did geoSurge raise, and what will it spend it on?

geoSurge raised an oversubscribed $12 million seed round, and it says it will spend the money on research and engineering teams, AI infrastructure and computing capacity, and development of Corpus Engineering. Pulse 2.0 and PYMNTS reported the round on July 3, 2026, and geoSurge describes it as the largest seed to date in the generative engine optimization category.

The size stands out against earlier seeds in the category. Profound announced a $3.5 million seed in 2024, Peec AI raised a €5.2 million seed in July 2025, and Brandlight raised a $5.75 million seed in April 2025. Forbes Argentina reported that geoSurge also runs its own cluster of Nvidia processors for research.

What is Corpus Engineering?

Corpus Engineering is geoSurge's name for shaping the text that AI models see, with the aim of changing how those models represent a brand. geoSurge says the method covers both what an AI system retrieves in real time and what it already holds from training.

The company separates two kinds of visibility. It calls what a model already knows from training parametric visibility, and what it looks up live dynamic visibility, and it argues that only the first is durable. Vigo told Forbes Argentina that as models become more capable they know more and search less, so the contest moves to how a model represents a brand over time.

That is geoSurge's thesis, and EPR has not found independent testing of it. Most other platforms, including Evertune, concentrate on measuring answers and citations. The wider discipline is explained in EPR's GEO guide, and a KDD 2024 paper by researchers at Princeton, Georgia Tech and IIT Delhi found that adding citations, statistics and quotations could raise visibility in generative engine responses by up to 40 percent, with results that varied by subject area.

Who founded geoSurge and who backs it?

Francisco Vigo and Jons Mostovojs co-founded geoSurge in London. Forbes Argentina reported that Vigo, an economist trained at the University of Buenos Aires, previously worked at Miniclip and Sensor Tower and led the data science team at a British fintech, and that he met Mostovojs, an engineer specializing in functional programming, in London.

Forbes Argentina also reported that geoSurge launched in January 2025 with an eight-slide pitch and signed with Passion Capital three weeks later, and that the company has more than 50 investors. The seed investors named in the announcement are AlbionVC, Play Ventures, Octopus Ventures, Celero Ventures, Boost Capital, Passion Capital and Tuesday Capital.

What has geoSurge published about Spanish-language AI answers?

geoSurge has published what Forbes Argentina calls the first AI visibility index in Spanish, based on 700 queries across seven industries and more than 14,000 answers from ChatGPT and Gemini. The company is also an official sponsor of the CMO LATAM Awards in Mexico and is starting a formal push into Latin America.

The index matters to communications teams because it treats Spanish as its own measurement problem, not a translation of English results. A brand that appears in English-language answers may be described differently, or not at all, when buyers ask the same question in Spanish, which is why the same question set should be run in each language.

How does geoSurge compare with other AI visibility startups?

geoSurge raised a larger seed than any of the other companies in the table. All figures are amounts announced by the companies.

The open question for geoSurge is whether a seed this large can be turned into dated evidence that training-level influence changes AI answers. The full category map is in EPR's Capital Behind AI Communications tracker.

What should PR and communications teams do with this news?

PR and communications teams should test separately what an AI assistant says with live search and what it says without it, because geoSurge's thesis depends on that difference. Where an assistant lets you turn web search off, ask the same question both ways and record the answers with the date.

1. Write a fixed set of buyer questions about your company and category, in each language you sell in.
2. Run each question with search on and, where the assistant allows it, with search off, and record the difference.
3. Keep your facts identical across your own site, press releases and the third-party pages that AI answers cite, since models are more likely to repeat sources that are cited often and agree with each other.
4. Ask any vendor selling training-level influence for dated before-and-after evidence before you pay for it. EPR's AI visibility audit method shows how to record results.