Welcome to Franc Talking.

🏔️ Speaking at Davos in the week, Sir Demis Hassabis stated he was ‘surprised’ that openAI had moved into ads so ‘early’ in their journey.

The Deepmind founder and creator of the 1994 smash hit game Theme Park (I’m always amazed at that - I loved this game) was speaking to Axios when he voiced his opinion.

Given Google’s ad-fuelled dominance, Hassabis was quick to clarify that “there is nothing wrong with ads”.

In a chatbot or personal-assistant environment, ads risk undermining neutrality. When users are asking for help, advice, or decisions, the moment they suspect commercial influence, confidence erodes fast.

This mirrors a point I discussed recently with Christian Ward from Yext on Franc Talking (episode dropping soon 👀) ads can’t feel random, intrusive, or irrelevant.

If you ask ChatGPT for the nearest sushi restaurant, you cannot be served an ad for Burger King.

Christian also hinted at what AI ads might eventually look like - more integrated, more contextual, more assistive. That still feels like the end state. For now, placing paid ads at the bottom of results looks like OpenAI testing the water, watching user behaviour closely before going further.

OpenAI’s CFO, Sarah Friar, pushed back, arguing that at ChatGPT’s scale, ads are inevitable. But that response neatly sidestepped the trust question and focused instead on timing.

None of this is especially surprising. Hassabis clearly took the opportunity to land a jab at OpenAI - and at Sam Altman, who has historically been outspokenly anti-ads in ChatGPT.

But the reality is this - burning through an estimated $115bn by 2029 forces hard choices. Monetising free users becomes unavoidable.

Ads represent a meaningful departure from OpenAI’s original mission — “to advance digital intelligence for the benefit of humanity, unconstrained by the need to generate financial return.” That feels a long time ago.

The honeymoon is done. And 2026 looks like a defining year for OpenAI - possibly even IPO territory.

A couple of weeks ago, Google’s Danny Sullivan said “we don’t want you to do that” when asked about breaking content into “bite-sized chunks” on the Search Off the Record podcast.

That response makes sense in context. Sullivan has seen this pattern play out repeatedly: SEOs latch onto a single comment, over-optimise in good faith, and then discover months later that the work was pointless or, worse, harmful. The disavow file is a perfect example of how well-intentioned optimisation can spiral into misuse.

At the same time, AI has dramatically lowered the barrier to producing content at scale. That reality has made discussions around formatting, structure and “chunking” far more polarised than they need to be.

The debate is often framed as black and white. It shouldn’t be.

If a page is currently a wall of text, improving its structure is going to help both humans and machines.

That means:

None of that is controversial. It improves readability, comprehension, and user trust - and it also happens to align well with how modern AI systems retrieve and summarise information.

Where things get lazy is when this advice is treated as a generic checklist. The real value comes from applying it based on what your business actually needs, not what Twitter or a podcast clip says you should do.

That’s why the more useful questions now look like this:

I’ve been working in SEO for over 20 years. I’ve seen plenty of “next big things” come and go - Google Glass, Google Buzz, Voice Search 1.0, and many more. I don’t sell tactics to capitalise on fads, and I don’t mislead people into believing fundamentals no longer matter.

That’s also why I don’t buy into the idea that GEO is “just SEO” - nor do I agree with the idea that the SEO playbook should be “thrown out the window”. That’s nonsense.

Many of the same principles still apply if you want to perform well in AI Overviews, AI Mode, or ChatGPT. What’s changed is the surface area. Authority, clarity, and evidence are now being tested in more places, in different ways.

For SEOs willing to understand that shift - rather than fight it or oversimplify it - the opportunity is enormous. And that’s something Christian Ward captures perfectly in the discussion below.

For further reading on this, I recommend checking out Duane Forrester newsletter from last week entitled “When Platforms Say ‘Don’t Optimize,’ Smart Teams Run Experiments”.

Earlier in the week I produced our first feature walkthrough and showcased how Obsero can help brands understand their content gaps and how smaller brands can out perform much larger brands with an investment in content.

Interested in seeing this in action? Mike and I would be happy to run you through it - book a demo here.

If you’re not already following us on LinkedIn, you should. I’m more vocal on my personal account, so connect there as well and add “Franc Talking reader 👋” in your note.

The best articles this week on AI Search and how it affects business, marketing and tech.