Welcome

Hello!

👋 Welcome to all new Franc Talking subscribers.

This edition breaks down how generative search (GEO) is reshaping traditional SEO and what marketers need to do to adapt.

Also, this is your weekly dose of business, marketing and tech updates. Diving in time! 🤿

If you're thinking about launching your own newsletter, I’ve set up a Franc Talking partner link — a great place to start.

If you’re a regular reader of this newsletter and haven’t yet filled in my super quick 30-second survey, I’d really appreciate it. 🙏 It helps me understand who’s tuning in, what you care about, and how I can shape future content (and partnerships) to make this even more valuable for you.

SEO's shift towards a GEO world

How GEO differs from SEO — and why that matters now

Acronyms aren’t new in marketing. But lately, there’s been a flare-up. Some SEOs are getting jumpy about the emergence of GEO (Generative Engine Optimisation) and AEO (Answer Engine Optimisation).

To some, GEO feels like a rebrand — a new wave of marketers jumping on a trend that overlaps heavily with what SEO has always done.

But here’s the thing: this tension reveals something deeper. A fear that the ground is shifting. And, spoiler alert — it is.

I’m not here to argue labels, although there is a very strong case to do so. I’m here to explain how search is evolving, what still works, and where attention needs to shift if you want to stay relevant.

What I cover in the newsletter: 👇

• Why GEO represents a structural shift in how search engines interpret and present content
• How traditional SEO techniques like digital PR, structured data and clear content hierarchy still influence generative visibility
• What actions to take now — from tracking AI visibility to influencing agent-friendly outputs

Early warning signs

We've been here before (kinda)

Twelve years ago, on my first day at Starcom, Google launched Hummingbird — the update that pivoted search toward understanding intent, not just keywords. I remember getting an email from a planner that read: “Welcome to Starcom. Just read this [article]...so....I guess SEO is dead?”

That was in 2013. Since then, I’ve worked across Wavemaker, IG Group, and Cazoo. Each year brought a new round of hand-wringing: “Is SEO being squeezed?” “Are zero-click searches rising?” “Do links still matter?”

If you’ve been in this industry for long enough, you get used to the existential threats. But this time, it’s not existential — it’s evolutionary. We’ve crossed into a new era. Not just in how people find information, but in how machines process, summarise, and serve it. And the term "search engine optimisation" doesn’t quite capture that anymore.

✏️ Beautiful templates

🚀 Powerful growth features

📈 Built-in analytics

💸 Monetisation tools

Changing channels

The machines are listening

Let’s talk usage. OpenAI’s ChatGPT reportedly processes around 2.5 billion prompts per day. Google? 14 billion searches — and counting.

Despite the huge growth, Google’s still dominant. But what’s changed is the expectation. People don’t just want links anymore — they want answers. They engage in conversation. They need tasks completed. Suggestions. Refined summaries.

We’re not optimising for search engines anymore — we’re optimising for outputs. For AI that consumes your site before a human even lays eyes on it.

Google’s AI Overviews are just the start. AI Mode is coming — and it will almost certainly replace AIOs within the year. The only thing holding it back now? Google’s need to safeguard its monetisation strategy. Welcome to the new world. Welcome to GEO.

Shifting views

Personalisation, prompts and visibility blind spots

LLMs don't show the same answers to every user. They personalise. Based on your history. Your behaviour. Your account. Even your mood.

That means one user’s AI summary might include your brand — and someone else’s won’t. You can’t just “track rankings” anymore, but you can and must have an understanding of how visible your brand is within AI Search, what citations are being used to power the answer and whether your content across your ecosystem is optimised for AI.

As AI Search grows, and it is far more quickly than expected, you need to simulate prompts based on your ICP (ideal customer profile). You need visibility tools built for generative search. You need to understand how your brand appears, or doesn’t, based on persona context.

Don't miss it

The opportunity hiding in plain sight

Yes, AI search is still small in traffic terms. But ChatGPT referrals grew 25.6% between May and June 2025. That’s pretty strong growth.

It reminds me of working with a betting firm back in the early 2010s. The CEO was reported to have asked: “Should we even bother with the Internet? Our punters are in the shops.”

The same logic is being applied now by brands writing off AI. It’s too early. Too messy. Too small. But ignore it, and you risk repeating the same mistake. The direction of travel is clear.

One of the biggest shifts in GEO is how authority is interpreted. Traditional SEO leaned heavily on links — now, it’s about mentions. Not just hyperlinked ones, but unlinked brand mentions across trusted sources: product reviews, forums, app stores, publisher sites, and comparison engines.

LLMs don’t need a clickable URL to surface your brand. They need to see you talked about, consistently, in the right places. In some cases, brand visibility — not backlinks — is the new currency of trust.

Live test

Testing ChatGPT's new agent feature

Less than two weeks ago, ChatGPT rolled out its new agent feature — and I gave it a test. I asked it to find me a UK car insurance quote for a BMW IX3, supplying my mileage and car details to get things moving. The agent immediately outlined what it needed — policy start date, DOB, claims history — and then began tasking itself.

First, it searched for UK insurance aggregators, picked CompareTheMarket, and tried loading their car insurance calculator. Blocked. Then it moved on to MoneySuperMarket — also blocked due to Cloudflare protection. Eventually, it attempted Confused.com, clicked through the initial prompts, and tried to proceed — but stalled out at the cookie/hCaptcha stage, requiring human input.

This wasn’t a hallucinated journey. Every step was logical, goal-oriented, and adaptive. The agent used fallback logic, navigated between URLs, selected the most prominent CTAs, and knew what data it needed to gather. It just couldn’t get past bot protection — yet.

This is early-stage A2A: agent-to-agent workflows where a user assigns a task, and the system tries to complete it. What stood out was how the agent prioritised certain domains — it chose mainstream aggregators, not niche brokers. That tells you something: it’s not just about being crawlable, but being recognisable and predictable to the agent logic stack.

From a marketing point of view, this shifts the game. These agents won’t show up in your referral reports. They’ll appear as:

• Unusual direct traffic to form-based pages
• Fast, multi-step navigation patterns from single sessions
• User agents with incomplete headers or JS-disabled sessions
• Unexpected brand mentions in generative interfaces without referral spikes

What you’re seeing is the beginning of something much bigger: machines learning how to shop, filter, and transact on your behalf. And if you’re not the site they can access and complete — you’re not even in the running.

This is very different to what is in the forever evolving SEO playbook and stating that "Good SEO will prevail" is just wrong.

Now what?

Six actions to take now

This isn’t just theory. If you’re working in organic, these are six things worth focusing on right now:

1. Audit your brand mentions across trusted sources:
   Check review sites, app stores, forums, publisher content, and comparison engines. Unlinked brand mentions carry weight — especially when models don’t rely on clickable links.
2. Structure your content for chunk-based retrieval:
   Use semantic headings, short paragraphs, clear sectioning, and schema (JSON-LD for FAQs, HowTo, Organisation). LLMs don’t read pages — they extract passages.
3. Rethink your performance metrics:
   Track branded search, impressions in AIOs, and downstream signals like direct or social conversions — not just click-throughs.
4. Run prompt simulations against your ICPs:
   Test how your brand appears in AI tools using real-world queries. Use ChatGPT, Perplexity, and Google’s AI Overview to stress test visibility.
5. Align PR and content with AI discoverability:
   Prioritise citations from domains likely to influence LLM training or retrieval — not just what passes as a "DA 80" link.
6. Start preparing for vector-based retrieval:
   LLMs retrieve answers based on meaning — not just keywords. Structure your content to be topically tight, semantically rich, and conceptually distinct. Think about how your site might look as a vector database — because that’s effectively what it is.

This isn’t a playbook rewrite — it’s a mindset shift. Visibility is still possible. But the game has changed.

Leaders in the space need to start talking like marketers and less like AI technicians, which can help the transition. “What does ‘vector-based retrieval’ mean for my KPIs?”
“Is my team set up to structure content for ‘chunk-based retrieval’?”

Final thoughts

The shift is already here

GEO is not a future trend - it’s already a live transition. AEO is part of that mix, but GEO better reflects the pivot we’re all experiencing: from optimising for search engine visibility (and clicks), to optimising for presence within generative, agent-led interfaces.

SEO has always evolved. But this isn’t just another evolution — it’s a restructuring of how discovery works. Machines are interpreting content, surfacing brands, and influencing decisions — often before a user even visits your site. And that’s where the difference lies.

The phrase “it’s just good SEO” underestimates the change. Traditional page-level tracking and attribution no longer capture the full picture. Influence is becoming harder to see — and easier to miss.

Most SEO professionals are well positioned to adapt — but the ones who thrive will be those who move early. That means understanding how LLMs surface information, working with data teams to redefine visibility metrics, and owning the narrative in the boardroom.

Don’t ignore the shift. Learn how generative systems decide, cite, and act. Understand where and how your brand appears across these interfaces. Discuss the implications with product and data teams. Build a POV you can take to leadership — and invest in visibility tools built for this new layer of discovery.

SEO was always a limiting label for what goes into the expertise. The game isn’t about search engine rankings anymore. It’s about brand visibility and relevance — even when no one clicks.