The Signal
Here is the new reality: Google still processes 14 billion searches daily while ChatGPT handles 37.5 million. The ratio is 373 to 1. But the trend line is unmistakable — and the businesses that show up in both places will have a compounding advantage. The question is not whether to optimise for AI. It is how to create content that works everywhere.
The Story: SEO is not dead — it evolved
Let me start with the numbers that matter. Google holds over 90% of the search market. That dominance is not going anywhere soon. But something important is shifting underneath it.
An Ahrefs study of 300,000 keywords found that when Google shows an AI Overview at the top of search results, the click-through rate on the first organic result drops by 34.5%. That means even if you rank number one, a third of your clicks may never arrive. The traffic you worked hard to earn is being absorbed by Google's own AI-generated answers.
At the same time, AI tools like ChatGPT, Perplexity, and Claude are becoming the starting point for a growing number of buyers. McKinsey research suggests approximately 50% of consumers now intentionally seek out AI-powered search when making purchasing decisions. They are not replacing Google — they are adding a new layer to how people find and evaluate businesses.
This is where Generative Engine Optimisation — GEO — comes in. In plain English, GEO means making your content so good, so clearly structured, and so genuinely useful that AI systems have no choice but to cite it when answering a relevant question. It is not a trick. It is not a hack. It is the logical extension of what good content marketing has always been about.
Why E-E-A-T matters more than keywords
Google introduced a framework called E-E-A-T — Experience, Expertise, Authoritativeness, and Trustworthiness. It started as guidance for the human quality raters who evaluate search results, but it has become the clearest signal for what both Google and AI systems prioritise.
Here is what each part means in practice:
Experience — Has the author actually done the thing they are writing about? A restaurant review from someone who ate there beats a summary written from other reviews.
Expertise — Does the author know this subject deeply? Credentials help, but demonstrated knowledge matters more. Show your working.
Authoritativeness — Is this source recognised by others in the field? Backlinks, citations, and mentions from trusted sources build authority over time.
Trustworthiness — Is the site transparent about who is behind it? Is the information accurate and well-sourced? Trust sits at the centre of the framework — without it, the other three do not count.
AI platforms evaluate these same signals. When ChatGPT or Perplexity chooses which sources to cite, they favour content tied to recognised entities, reliable sources, and verifiable claims. The content that wins is the content that earns trust across both human readers and AI systems.
Why human-written content still wins
This is the part that surprises people. In a world flooded with AI-generated articles, content written by real humans with real experience is becoming more valuable, not less. AI systems are trained to recognise quality, and they can increasingly distinguish between genuinely useful content and the generic, keyword-stuffed articles that AI writing tools produce at scale.
A niche business blog written by someone who genuinely understands their customers' problems can outrank a corporate giant's content library — because it answers real questions with real specificity. AI retrieval systems notice when content includes unique data, original perspectives, and insights that cannot be found elsewhere. That is your competitive advantage.
The principles we share with clients
When we advise clients at LF Labs on content strategy, the principles are simple: write from genuine experience, structure your content so both humans and machines can parse it easily, answer specific questions directly, and update your content regularly. No shortcuts, no gaming the system. Just consistent, useful work that compounds over time.
The Operator's Toolkit: The content structure checklist
Print this out. Use it every time you publish a page or blog post.
Clear headings with descriptive labels. Not clever, not vague. If someone scans only your headings, they should understand exactly what each section covers. AI systems parse headings to determine relevance.
Direct answers within the first paragraph of each section. Do not bury the point. State it clearly, then expand. AI systems pull from the first sentences when generating answers.
FAQ sections with genuine questions. Use the actual questions your customers ask — from support tickets, sales calls, and emails. Not the questions you wish they asked.
Author attribution with real credentials. Put a name, title, and brief bio on every piece of content. This is a direct E-E-A-T signal that both Google and AI systems use to evaluate trustworthiness.
Regular content updates. AI retrieval systems favour fresh content. Set a quarterly review schedule for your top-performing pages. Update statistics, add new examples, and note the "last updated" date visibly.
Structured data and schema markup. Implement Article, FAQ, Organisation, and HowTo schema where relevant. This helps AI systems understand and categorise your content correctly. If this sounds technical, ask your developer — it is a straightforward implementation.
At least one unique insight per piece. Every article should contain a data point, perspective, or piece of analysis that cannot be found elsewhere. This is what makes AI systems cite you instead of the other ten articles covering the same topic.
The Radar: Three things worth knowing this week
1. Content volume drives AI visibility faster than you might think. Data from Brandi AI shows that brands producing 12 or more optimised content pieces monthly see AI visibility gains up to 200 times faster than those producing only four. Consistency is not optional — it is the multiplier.
2. Black hat LLM SEO is already a problem. Search Engine Journal reports that bad actors are creating microsites, fabricating author personas, and poisoning AI training data to manipulate how AI systems recommend brands. Multi-billion-dollar companies are not immune — some are building networks of sites specifically to game AI citations. The good news: best practices still win. Quality and authenticity are harder to fake at scale.
3. IndexNow protocol: get your content in front of AI faster. IndexNow lets you notify search engines the moment you publish, instead of waiting days for crawlers to find your new content. It takes about 30 minutes to set up, and participating search engines index notified URLs five to ten times faster. If timing matters for your content strategy, this is a quick win.
From the Field
Here is my question for you: Take a look at your website. Which page do you think AI would cite if someone asked about your industry — and which page would it ignore completely?
Hit reply and tell me. I am genuinely curious what you think, and I will share the most interesting answers in a future edition.
Until next time,
Francois