GEO Myths: Separating Evidence from Hype in AI Search Optimization

GEO Myths: Separating Evidence from Hype in AI Search Optimization

Generative Engine Optimization (GEO) is bringing a wave of new recommendations — some helpful, many unproven. Search Engine Land contributor Philipp Götza’s January 19, 2026 piece highlights how quick adoption of untested tactics can mislead marketers. As Götza warns, “We fall for bad GEO and SEO advice because of ignorance, stupidity, cognitive biases, and black-and-white thinking.”

Why GEO advice spreads — and why that’s a problem

Götza’s core point is that many popular GEO claims sit low on the ladder of evidence: statements repeated until they feel like facts, but unsupported by robust data. The result is a fast-moving cycle of attention-grabbing claims and “workslop” — content that looks authoritative but collapses on scrutiny. This is particularly risky when teams shift resources to follow definitive-sounding directives without validation.

Three pervasive GEO recommendations — what the evidence shows

1. llms.txt will make AIs cite your site

The llms.txt proposal suggests adding a concise, LLM-friendly file at /llms.txt and publishing .md versions of pages to help models find authoritative content. While llms.txt has merit as a standardization effort, Götza notes there is currently “no evidence — or proof — that an llms.txt meaningfully influences your AI presence.” Implementing llms.txt could be useful for certain technical doc sites or APIs, but for most publishers it’s premature to treat the file as a visibility shortcut.

2. Schema markup guarantees AI citations

Schema remains a hygiene factor for search visibility and rich results. However, Götza cautions that there’s no conclusive proof AI chatbots use schema as a direct citation signal. Correlations exist between schema usage and AI visibility, but rival explanations are plausible. Implement schema for its search benefits and clarity for human users — not because it will automatically trigger AI citations.

3. Freshness always wins

Among the three recommendations, freshness has the strongest empirical backing. Large-scale studies show AI assistants often prefer newer sources: “Compared to traditional search results, AI assistants prefer citing fresher content,” a research summary from Ahrefs found. Updating genuinely useful content and surfacing clearly dated updates increases the chance of being cited by AI systems that favor recency for certain queries.

Actionable guidance for site owners and SEOs

Götza offers a practical framework — evaluate claims using a ladder of misinference (statement → fact → data → evidence → proof) — and we translate that into steps you can apply today:

• Vet recommendations before committing resources: Run small tests or A/B experiments to measure actual impact on AI visibility or organic metrics.
• Prioritize content freshness where it matters: For queries where timeliness affects accuracy (news, product availability, health, policies), update content and surface last-modified dates consistently in-page, in schema, and in sitemaps.
• Keep schema as standard SEO hygiene: Implement relevant schema types accurately to support rich results and better indexing, but don’t expect immediate GEO inclusion from markup alone.
• Track crawl and access behavior: If you experiment with llms.txt, monitor logs to see whether AI agents actually fetch the file and whether crawl volumes change meaningfully.
• Resist AI-only summaries: Götza explicitly cautions against relying on AI for summarization; human review reduces the spread of misinterpretation and prevents echo-chamber amplification.

Implications for strategy and resourcing

Organizations should avoid treating GEO recommendations as binary rules. Instead, fold GEO experiments into existing SEO and content workflows. Invest in measurement, maintain editorial rigor, and favor sustained improvements over attention-grabbing hacks. In many cases, the highest ROI will come from publishing clearly sourced, high-quality content and improving user experience rather than chasing unproven standards.

Final thoughts and attribution

Philipp Götza’s piece is a timely reminder to substitute curiosity and testing for dogma. As he concludes, pause before you believe and test before you scale. Complement that caution with evidence-based research: Ahrefs’ analysis found clear freshness effects that teams can act on, but also cautioned that freshness isn’t a silver bullet. “Compared to traditional search results, AI assistants prefer citing fresher content,” Ahrefs wrote, while noting many other factors remain important.

For seoteric.com readers: treat llms.txt and other emerging standards as potential tools, not guarantees. Use the ladder of misinference, run small experiments, and prioritize genuinely useful updates where they matter.

Original article: GEO myths: This article may contain lies — Philipp Götza, Search Engine Land (January 19, 2026).

Additional sources: llms.txt proposal, Ahrefs — Do AI assistants prefer to cite fresh content?