Opinion
From SEO to GEO: Fighting for visibility in the age of AI
"Scaling GEO/AEO requires analytics for AI visibility, standardized metrics for recall, and ingestion pipelines feeding structured, brand-owned data into models," writes Sharon Kinory, an Associate at StageOne Ventures.
For over 20 years, Search Engine Optimization dictated online visibility. We mastered algorithms from AltaVista to Google, learned how to rank, and built industries on the mechanics of discoverability.
That playbook is broken. Search is no longer a list of links, its conversation based. Generative AI doesn’t point to sources; it delivers synthesized answers. Google’s AI Overviews summarize the web at the top of results. Perplexity’s Comet and other AI browsers go further, embedding summaries, follow-ups, and fact-checks directly into browsing.
Together, they don’t just change how we find information, they’re rewriting who gets seen, trusted, and rewarded online. The result? Less clicking, less traffic, and fewer opportunities for creators, small businesses, and publishers.
According to Similarweb, publisher search traffic has dropped 26% since AI Overviews rolled out roughly one in four visits gone. Meanwhile, referrals from ChatGPT have surged 25x, with Reuters and the New York Post leading among AI-driven sources. Brands lose traffic, ad impressions, and subscriptions. Monday.com’s stock plunged 26% after citing AI Overviews as a driver of slowing traffic. At the same time, competition for control is intensifying Perplexity even made a $34.5B cash offer to acquire Google Chrome, underscoring how critical the interface has become.
The New Visibility Crisis
Traditional SEO had rules you could test and rankings you could track. LLMs offer none of that. Their outputs are opaque, probabilistic, and constantly shifting.
Two new fronts are emerging.
The first is Generative Engine Optimization (GEO) and Answer Engine Optimization (AEO), which aim to influence how models like GPT, Claude, or Perplexity mention and represent your brand. Unlike search, these engines deliver synthesized answers, fewer clicks, less control, and no transparent system. The challenge is to ensure your presence in AI outputs, even inside a black box.
The second front is the rise of AI-driven browsers such as Perplexity’s Comet or Dia, with bigger players soon to follow. These don’t just display pages, they reshape interaction by summarizing videos, adding context, or fact-checking in real time. Even when users reach your site, their first touchpoint may be the AI’s condensed version. Optimization shifts from keyword targeting to influencing how AI layers summarize, prioritize, and frame your material.
Four Hard Problems
- Zero observability: No rankings, no CTRs. Mentions without clicks are common, benchmarks remain experimental.
- Unclear pathways: Visibility depends on unpredictable training data and retrieval pipelines.
- Vulnerabilities: Models can misrepresent or be manipulated.
- RAG complexity: Retrieval-Augmented Generation (RAG) changes the rules. Instead of relying only on what a model knows, it pulls live data. That means two demands:
- Findability: content must surface in the retrieval layer.
- AI-readability: clear structure, headings, and metadata so models capture the core message.
Even then, there’s no guarantee: insights may be condensed or merged with competitors’ data. Marketing and technical teams must align content must not just rank but be structured so AI can preserve its meaning.
Why GEO and AI Browsers Matter Together
GEO determines if you’re mentioned in AI-driven answers. AI browser optimization determines how you’re represented.
The implications are clear:
- Traffic becomes secondary presence comes first
- Priorities shift - from pleasing human readers to influencing AI intermediaries
- Defensibility matters - those who master AI framing and brand protection will win
How do you market in a world where the gatekeeper isn’t human, and what does winning even look like? The old playbook of rankings, ad impressions, and predictable funnels no longer applies. Visibility is now negotiated with AI systems that decide what’s seen, how it’s framed, and whether it’s trusted. The race has already begun.
Execution Layer
The goal is to live in an AI’s internal “mental model” of your brand. Early movers like Profound track mentions across AI outputs, XFunnel simulates buyer queries to test recall, and Brandlight maps visibility against competitors on ChatGPT, Gemini, and Perplexity. Still, no one has cracked how to reliably influence an LLM’s brand narrative. For now, it’s frontier work: prompt testing, content structuring, and iteration until you earn AI citations.
Infrastructure Layer
Scaling GEO/AEO requires analytics for AI visibility, standardized metrics for recall, and ingestion pipelines feeding structured, brand-owned data into models. On top of that is monetization: how do you turn AI references into revenue when users don’t click through? Unlike SEO’s ad-and-affiliate rails, GEO will need new commercial rails: pay-per-crawl models, verified brand APIs, and attribution systems.
This is one of the most consequential second-order effects of the AI revolution that began in 2022. The first wave-built models; this wave builds the infrastructure and rails. Rewards will be outsized for early entrants who define systems in data, measurement, or monetization. But volatility is inevitable: rules are undefined, platform power concentrated, and the market will keep shifting as adoption deepens. It’s a high-risk, high leverage bet on the foundations of a new “attention economy.”
Sharon Kinory is an Associate at StageOne Ventures.
