When Benjamin Houy recently announced that he was shutting down Lorelight, the generative AI visibility-tool he built to monitor brand mentions in ChatGPT, Claude and Perplexity, he rattled more than a few industry expectations. His key thesis: there is no sustainable “GEO strategy” separate from brand building.

"There is no sustainable “GEO strategy” separate from brand building."

What is GEO, really?

At its core, GEO is the idea that you can optimize for being cited or recognized inside AI-generated outputs, not just for ranking in search engine result pages. It asks: “How do I become a trusted source that an LLM (large language model) will draw from or mention by name?”

But what many early GEO thinking omitted is that LLMs draw heavily from the same content ecosystems we’ve been optimizing for in SEO, PR, content marketing, and brand reputation. The distance between “SEO + brand” and “GEO success” is smaller than many early proponents implied.

A new academic paper, Generative Engine Optimization: How to Dominate AI Search, backs this: AI systems show strong bias toward “earned media”, citations from third-party, authoritative sources, over mere owned sites or social media.

Why most standalone GEO tools fail

Houy’s decision reveals a few deeper truths about what went wrong with treating GEO as a separate layer you can bolt on:

Insight without leverage: Many users of GEO dashboards would see “mentions in ChatGPT answers” or “visibility score changes,” but those insights rarely translated into new strategy. The underlying levers were the same as always.
Overemphasis on measurement, underemphasis on impact: You can track a lot, but tracking alone doesn’t make you more authoritative or more cited.
Dependence on brand strength: Without a base of credibility, optimized content, and external recognition, many brands were invisible regardless of how “GEO-friendly” their pages were.
Redundancy: As AI systems evolve, many GEO signals (like structured question/answer formats, clarity, schema) get folded into standard SEO considerations.

Hence, the smartest move was to shut the tool down and refocus on building what actually moves the needle: brand equity, content depth, and broad visibility.

How to think about GEO within brand strategy

Rather than fighting GEO as a separate domain, treat generative visibility as an outcome, not a silo. Here’s a framework for that:

Pillar	What to Focus On	Why It Helps
Earned media & citations	Secure mentions in trusted third-party outlets: industry journals, forums, respected media	AI models favor external validation more than self-promotional content.
Content clarity + extractability	Structure content for clarity (FAQ blocks, clean headings, self-contained definitions, numbered steps)	Makes it easier for AI to “quote” or summarize correctly.
Topical depth & authority	Cover a topic comprehensively, linking between subtopics, updating content, showing cases, research	Deep topical coverage signals authority, more margin for AI to surface you.
Broad distribution	Share in forums, newsletters, podcasts, guest posts, academic sites, stack exchanges	Increases the chance AI “sees” your work in more contexts.
Technical hygiene	Fast loading, schema data, accessibility, correct metadata, semantic markup	Helps crawlers and AI preprocessors parse your content with less noise.
Measurement via signal lift	Brand search lift, branded traffic, AI citation audits (manually or via tools)	Focus on signal improvements over vanity metrics.

This is not radically new. It’s remixing what good content, link, and brand strategies have always done, but with one shift: metrics of success expand. You don’t just measure SERP position, but also “Are we being mentioned in AI answers that matter?”

When GEO thinking is useful

That isn’t to say GEO is meaningless. In these conditions, GEO frameworks add value:

Newer brands with low domain strength: They can use structured, AI-friendly content early to gain traction in AI ecosystems.
Edge cases / niches: In verticals where AI answers are less saturated, intelligent structuring and early presence can lead to disproportionate visibility.
Experimentation & feedback loops: Use small bets to test what questions AI picks up from your content, refine accordingly.

But GEO should not be a “magic lever” you depend on, it’s a magnifier of strength, not a substitute.

A twist: The reputation floor

One angle that I believe deserves more weight is the concept of a reputation floor for GEO. If your brand (or author identity) doesn’t have minimal credibility, an “empty brand", AI is less likely to trust or cite you, regardless of how well your content is structured.

So a practical startup or independent creator should invest first in building that floor:

Author identity - make clear who writes, with credentials, experience, transparency.
Anchor citations - get quoted by peer sites, experts, forums.
Community engagement - win recognition in your niche, create conversations.
Consistent content over time - the long game still wins.

Once you have that foundation, GEO-aware optimizations become more powerful, not just supplemental.

Conclusion: GEO is not a detour, it’s the vantage point you aim toward

Lorelight’s shutdown is a lesson: trying to treat GEO as a separate productized layer leads to diminishing returns. The real lift comes from building brand, authority, and content that existing systems already reward and folding generative visibility into that.

If you build brand first, GEO becomes not a new frontier to conquer, but a new horizon that extends the visibility of your work.