AI-driven search is fracturing fast. ChatGPT’s reasoning modes pull from different source pools. Google is building a new open standard for how AI agents consume knowledge. And a staggering 62% of AI citations never actually mention the brand they link to. We’re tracking these shifts daily for our clients, and the picture is clear: generative engine optimisation (GEO) now demands a fundamentally different playbook from traditional SEO. Here’s what matters right now.

Key Takeaways

• Only 25% of cited sources overlap between ChatGPT’s Thinking mode and Instant mode — treating ChatGPT as one system is a strategic error.
• 62% of AI citations across ChatGPT, Gemini, and Google AI Overviews are “ghost citations” that link to a domain but never mention the brand by name.
• Google’s new Open Knowledge Format (OKF) sets a standard for how AI agents read and use structured knowledge.
• Competitor keyword analysis must now span traditional search, AI Search, and web conversations to close visibility gaps.
• Improving E-E-A-T signals and targeting low-competition keywords remain the most reliable levers for ranking in both organic results and AI Overviews.

ChatGPT’s Reasoning Modes Cite Different Sources — Plan Accordingly

Most brands assume ChatGPT is a monolith. It isn’t. Our team was struck by research showing that only 25% of cited sources overlap between ChatGPT’s Thinking mode and Instant mode. Each mode favours different content types, different domains, and can even recommend different brands for the exact same query.

For our clients, this means we can’t optimise for “ChatGPT” as a single channel. We’re now auditing visibility separately across reasoning modes, checking which content surfaces where, and adjusting content depth and format to match each mode’s preferences. If your agency isn’t doing this, you’re likely invisible in at least one mode.

Ghost Citations: Getting Linked Without Getting Named

Being cited by an AI engine sounds like a win. Often, it’s hollow. A joint study with Kevin Indig analysed 3,981 domain appearances and found that 62% of AI citations don’t lead to brand mentions across ChatGPT, Gemini, Google AI Overviews, and Google AI Mode. These “ghost citations” link to your page but never say your brand name in the AI-generated answer.

We’re tackling this head-on. Our content strategies now embed brand names, product names, and service descriptors directly into the most citation-worthy sections of every page — typically the opening paragraph, structured data, and any summary or definition blocks. The goal: make it impossible for an AI engine to cite the content without also surfacing the brand.

Google’s Open Knowledge Format Sets the Rules for AI Agents

Google has released what it calls the Open Knowledge Format (OKF), a standardised way for AI agents to consume and interpret structured knowledge. As detailed in this breakdown of Google’s Open Knowledge Format, OKF aims to give AI systems a consistent schema for reading factual content.

For our team, this is an immediate action item. We’re reviewing client sites to ensure structured data, knowledge panels, and entity markup align with OKF expectations. Early adopters will have a clear advantage as AI agents increasingly rely on standardised formats to determine which sources are authoritative.

Competitor Keyword Analysis Now Spans Three Channels

Traditional keyword gap analysis isn’t enough. We’re now running competitor research across organic search, AI-generated results, and broader web conversations — forums, Reddit threads, social mentions. This multi-channel approach, outlined in this guide to finding competitor keywords and closing visibility gaps, lets us spot where rivals are being cited by AI engines and where our clients are missing entirely.

E-E-A-T and Low-Competition Keywords Still Drive Results

Amid all the AI disruption, the fundamentals hold. Strengthening E-E-A-T signals — real expertise, firsthand experience, clear authorship — remains the most effective way to rank in both traditional results and AI Overviews. We’re pairing this with systematic targeting of low-competition, high-intent keywords, a strategy validated by these nine proven methods for improving SEO rankings and traffic. For our clients, this combination consistently delivers measurable gains regardless of how AI search evolves.

The GEO landscape is splitting into distinct surfaces, each with its own rules. Agencies and in-house teams that treat AI search as a single channel, or ignore ghost citations, will lose ground quickly. We’re rebuilding our optimisation workflows around these realities — and advising every client to do the same.

Frequently Asked Questions

What are ghost citations in AI search?

Ghost citations occur when an AI engine like ChatGPT or Gemini links to your page but never mentions your brand name in its generated answer. Research shows this happens in 62% of AI citations, meaning brands miss out on recognition even when their content is used as a source.

How do web designers optimise content for different ChatGPT reasoning modes?

You need to audit which of your pages surface in Thinking mode versus Instant mode separately, since only 25% of sources overlap. Adjust content depth, format, and structure to match each mode’s preferences — longer, more analytical content tends to perform differently from concise, direct answers.

What is Google’s Open Knowledge Format and why does it matter for SEO?

Open Knowledge Format (OKF) is Google’s new standard for structuring knowledge so AI agents can read and use it consistently. Sites that align their structured data and entity markup with OKF early will be better positioned to appear as authoritative sources in AI-generated results.

Why does E-E-A-T still matter for AI Overviews?

AI engines prioritise sources that demonstrate genuine expertise, experience, authority, and trustworthiness when selecting which content to cite. Strengthening these signals on your site increases the likelihood of being surfaced in both traditional organic results and AI-generated answers.