Do people still Google? How to Optimise for Generative Search & LLMs

Introduction

For years, SEO was a game of keywords, backlinks, and keeping the Google gods happy. But let’s be honest with overselves: is that really what search looks like now?

Today, your client types a question into Google and boom, an AI Overview gives them the answer before they even glance at the first organic result. Or maybe they ask ChatGPT, Bing’s Copilot or Perplexity and get a neat little summary with sources.

If you're an agency targeting growth for your client, here’s the real question you should be asking yourself:

So, If search is generating answers for people, not just links, how do we make sure our clients’ names show up in those answers?

That’s what this post is about.

---

What Are Generative Search and LLMs?

First, let’s break down the buzzwords.

• Large Language Models (LLMs): These are the brains behind the curtain. Trained on oceans of text, they don’t just match keywords; they understand meaning and context. (I'm looking at GPT-4, Claude, Gemini, etc.)
• Generative Search: Instead of “ten blue links”, users now get a written answer. Powered by an LLM. Right at the top of search. Google’s AI Overviews are the poster child for this.
• RAG (Retrieval-Augmented Generation): The LLM doesn’t just rely on what it was trained on. It also fetches fresh sources in real time to improve accuracy.

As an agency, the question to ask is: if LLMs are the ones writing answers, how do we become the sources they pull from?

---

Why This Matters for Agencies

Remember when “position zero” (featured snippets) felt like the holy grail? Now imagine if those snippets covered the entire screen. That’s where we are with AI answers.

• Risk: Organic clicks shrink when users don’t need to click through.
• Reward: If your client’s content is cited or used, they become the brand in the spotlight.

So ask yourself: what happens when your client’s competitor is quoted in an AI answer and they’re not?

---

Traditional SEO vs Generative SEO

Old Playbook	New Reality
“Which keyword should we target?”	“Which question will AI need answered?”
“How many backlinks can we build?”	“Does this content show topical authority?”
“Are we ranking top 3?”	“Are we cited inside AI summaries?”

See the shift? It’s not that old tactics are dead — it’s that they’re not enough.

---

Strategies to Optimise for Generative Search & LLMs

1. Build Semantic Authority with Content Clusters

Ask yourself: does our site look like an expert, or just a dabbler?

• Create pillar pages on broad topics.
• Support with cluster articles answering niche, long-tail questions.
• Interlink everything.

This helps LLMs recognise your client as a thought leader, not a one-off.

---

2. Optimise for Questions, Not Just Keywords

The AI is built to respond to conversational queries. So here's the key: are we writing content in the same language our audience asks?

• Use tools like People Also Ask to see actual phrasing.
• Bake FAQs into blog posts and service pages.
• Write “What is...?” and “How do I...?” answers in plain English.

If your copy looks like it’s answering the exact question, you’ve got a shot at being quoted.

---

3. Structure Content for Easy Extraction

LLMs are like busy interns. They skim. So you have to ask: is our content skimmable enough for an AI to grab the right line?

• Short paragraphs.
• Clear headings.
• Bullet lists and tables.
• TL;DR sections or one-sentence takeaways.

Think of every blog post as having “quotable nuggets” an AI can lift.

---

4. Double Down on Authority & Trust

Why should an AI trust your client’s site over a competitor’s? Because you prove credibility.

• Add author bios with expertise.
• Reference credible third-party sources.
• Include original stats, case studies, or research.

Ask yourself, does this piece of content look like it could be cited in a research paper? If not, beef it up.

---

5. Keep Content Fresh

AI models lean towards current, relevant sources.

• Refresh cornerstone pages every 6–12 months.
• Update stats, screenshots, and product details.
• Add new Q&A sections when industry trends shift.

Ask again, if someone read this today, would they believe it was written last week or five years ago?

---

6. Schema & Structured Data

Schema is like leaving breadcrumbs for the AI.

• FAQ schema.
• Article schema.
• HowTo schema.

It’s not magic, but it’s a signal. Another question for you, have we done everything possible to make our content machine-readable?

---

7. Monitor “AI Visibility”

The hardest part right now is measurement. But you can’t ignore it.

• Track mentions in AI Overviews where possible.
• Monitor brand citations with tools like Ahrefs Alerts.
• Talk to clients. Leads might come from “I saw you in a Google AI answer.”

Have a basic system in place to track AI-driven visibility.

---

What Not to Waste Time On

There are already distractions in the space. For example:

• LLMS.txt: Google has said they’re not using it.
• Keyword stuffing: Won’t fool an AI.
• Thin content: If it doesn’t help a human, it won’t help in generative search.

Ask yourself: are we chasing shiny objects, or building lasting authority?

---

Challenges Ahead

Let’s be real, this is new territory.

• The rules aren’t fully transparent.
• Click-through rates may drop, even if visibility rises.
• Platforms will change how they use sources.

But if you’re focusing on the right things, what do our clients’ audiences actually want, and how do we provide it better than anyone else? You’ll stay ahead.

---

Conclusion

Generative search and LLMs aren’t killing SEO. They’re rewriting it.

So as an agency leader or copywriter, the question you should be asking yourself today isn’t “what keywords do we target?” but:

What answers do we want AI to give? How do we make sure they come from us?

If you can answer that, you’re not just optimising for search. You’re future-proofing in a world where AI is the new front door to the internet.