Generative Engine Optimisation is rapidly becoming a critical strategic requirement across industries, yet most organisations lack the measurement frameworks, organisational structures, and specialist expertise to pursue it effectively on their own. ChatGPT, Gemini, Claude, and AI-powered search features like Google's AI Overviews are now the first point of contact for a growing share of product and service discovery.

Users ask questions in natural language and receive synthesised answers, often without ever viewing a traditional search results page. AI-referred sessions jumped 527% year-over-year in the first five months of 2025 — and most organisations are not yet measuring this shift, let alone responding to it.

This shift is what practitioners now refer to as the GEO challenge: Generative Engine Optimisation, the discipline of ensuring that a brand, product, or narrative is accurately and favourably represented within AI-generated responses. GEO is the natural evolution of traditional SEO. The distinction matters because ChatGPT cites an average of 5–8 sources per response — far fewer than Google's organic results — while Perplexity cites significantly more but with almost no overlap. It is, structurally, a winner-takes-all environment, and the positions are beginning to be established now.

As Marta Alvarez explains: “GEO doesn’t replace SEO; it amplifies it. Companies in 2026 need an AI visibility strategy. Organisations that postpone adoption risk creating a structural competitive gap that late movers will find increasingly difficult to close. Traffic from AI-driven search is still a relatively small share but it's a future-proofing strategy. Companies need to prepare for an AI-first era."

The challenge is mostly structural because GEO sits at the intersection of content strategy, technical SEO, data infrastructure, brand authority, PR, and analytics — each of which has historically operated in its own silo, with its own metrics and reporting lines. GEO requires all of them to work in coordination, measured against a unified framework, at a moment when the measurement tools themselves are still being built.

The measurement problem is fundamental – not cosmetic

The most fundamental obstacle companies face with GEO is the absence of a coherent measurement framework. Unlike traditional search, where you have rankings, clicks, and impressions, brands optimising for AI visibility are largely blind — without reliable data on whether they're being cited, how often, or in what context. This makes it hard to measure ROI, prioritise investment, or even build internal confidence in GEO initiatives.

In traditional SEO, the feedback loop was hard-won but ultimately legible through rank position, CTR, impressions, bounce rate. All of it flowed through GA4, Search Console, or your CDP. GEO flips this entirely — enterprises now face the challenge of measuring visibility when users never touch the SERP or even click through to a website. That means your traditional conversion funnel is structurally blind to a growing portion of your brand surface area.

AI & Data Science Director Marcello Cacciato frames the problem around two compounding factors: "The first is simply maturity: GEO is new, and the measurement conventions are still forming. The second is more structural, and harder to resolve. This is what I call the black box problem. AI models don't disclose exactly how they capture and weigh information from websites. Some things we know, some we infer, but there's intellectual property involved, and that limits our inference power. Because of this, we shouldn't claim that the community has already agreed on the ultimate GEO KPIs. What Metyis has done instead is to build a directional scoring approach — a score from zero to 100 that tells you whether you're far behind, moving in the right direction, or covering the essentials. It's not a perfect KPI, but it gives direction."

GEO also introduces what Marta Alvarez describes as zero-click funnels: users receive an answer within the AI interface and never visit the brand's website at all. As she explains: "Traffic volumes may decline, but the traffic you do get is likely to be very high intent. Users may already have made a decision before even visiting your site. This breaks the old marketing model, attribution and channel tracking. The KPIs we used to optimise for — clicks, impressions, rankings — are no longer sufficient. New metrics are still being defined, and most teams aren't familiar with them yet. That's why dashboards alone aren't enough; teams need interpretation and insight."

Content authority as a data asset

What most companies miss is that LLMs only cite 2–7 domains on average per response, far fewer than Google's 10 blue links. AI engines rely on Retrieval-Augmented Generation (RAG), augmenting generative models with external documents retrieved in real time. The implication is that structured, citable, entity-rich content becomes a scarce asset rather than a volume play. Nearly 80% of top news publishers now block at least one AI training crawler, creating a content scarcity dynamic where brands that make their content AI-accessible gain an outsized advantage.

Marcello Cacciato identifies content authority as a core strategic lever, and one that is widely misunderstood: "AI assigns weight to different sources. If you're a brand, you want AI to pull your story from your own authoritative sources — not from a random forum or influencer post. That's content authority." If a company does not ensure that its own properties are the highest-authority sources for its story, AI models will fill that gap from wherever they can. Third-party sources such as forums, review platforms, Q&A communities, and social media become the default inputs. The resulting narrative may be accurate, or it may not be, but it will not be the one the brand would have chosen.

Third-party presence therefore becomes a strategic component of GEO in a way that goes well beyond traditional PR. Marta Alvarez is direct about the implications: "Unlike traditional SEO, AI pulls from many sources you don't control — YouTube, Reddit, reviews, Q&A platforms. Third-party mentions carry a lot of weight. Controlling your brand narrative means not only optimising your own website, but also placing accurate, authoritative content on external platforms. This touches PR as much as SEO. If you don't do this, AI will still tell a story about your brand but it may not be the one you want."

First-party data: from privacy asset to discovery advantage

Most companies invested in first-party data to survive cookie deprecation. GEO adds a second strategic motive: the brands that feed proprietary, structured, first-party signals into AI platforms directly — think ChatGPT's merchant feed programme, or Amazon's Rufus — will have a citation advantage that pure content optimisation cannot replicate. Your CRM, product catalogue, and inventory data become GEO assets. Most data strategies weren't written with that use case in mind.

Marcello Cacciato explains how this plays out in practice: "First-party data becomes even more important in GEO. Your CRM data can tell you who your core customers are and what questions they're likely to ask AI. You can then structure your website content around those questions and answers. That's GEO in practice: using insights you already have to become visible in AI engines."

The companies that will win in GEO aren't just publishing better content. They're operationalising customer and product data as part of how AI systems discover and retrieve information. That requires data engineering, governance, and a strategic frame that most marketing teams don't own, and most data teams haven't been briefed on. This is where the real constraint emerges: GEO is not just a content or optimisation problem, but a question of how capabilities are structured across the organisation.

Organisational structure as a constraint

The way marketing organisations are currently structured was not designed for what GEO requires, and Marcello Cacciato explains why: "Traditionally, marketing teams are organised into separate silos — branding, data, SEO. GEO cuts across all of them and requires a unified approach and unified measurement, which is something organisations haven't needed at this level before. That's also why there aren't many GEO experts: you need expertise across many disciplines." Marta Alvarez reinforces this point: "Finding a single team or person that covers everything — from content strategy, SEO, marketing, PR, tech, data, to infrastructure — is extremely rare. Successful GEO requires a new operating model and a level of coordination that most organisations aren't structured for today."

Traditional SEO teams are being forced to upskill quickly, often without clear playbooks, while still maintaining traditional search performance. The broader GEO talent pool remains limited and the vendor ecosystem is still immature. As a result, many companies encounter tools that promise measurement without execution guidance, or strategic advice without reliable visibility data. Without both, it becomes difficult to connect GEO activity to measurable business impact.

The structural case for acting now

For organisations trying to make sense of GEO, it is important to understand that the urgency is real but it is not a reason to panic. There is no need to reallocate entire budgets or overhaul existing efforts, but in 2026 organisations should at least have an AI visibility strategy in place.

Delaying will create a gap that will be structurally very hard for late movers to close in a channel that will only continue to scale. As Marta Alvarez puts it: "The biggest blocker for most companies is simply not knowing where to start. The key is to get started, understand your baseline, and take some steps. Late movers will struggle to catch up."

Marcello Cacciato offers a similarly measured but firm assessment of the trajectory: "The technology will keep evolving, but the fundamental change has already happened: large language models are now a primary discovery layer for the web. That's not going away. Companies need to accept that and adapt."