Beyond Organic: The Archive
Beyond Organic is my newsletter on international SEO and AI search. This page collects every past issue. If you'd rather get the next one in your inbox, subscribe here.
AI Overviews launched in France
Google launched AI Overviews in France last week, two months ahead of the deadline it had given publishers and with no signed remuneration agreement in place. Here's what the data from other markets shows, and the one thing worth doing before your Q4 numbers move for reasons that have nothing to do with your team.
Hey,
Google launched AI Overviews in France last Wednesday. If you don't work the French market, you probably didn't notice, and there's no real reason you should have. France is market one hundred and something. By international standards the story is over before it starts.
That reading is roughly right. It's also why this one is easy to get wrong.
The asymmetry I watched from both sides
In French SEO circles, the launch produced seven days of noise. Studies, trigger rates, turnkey solutions, people announcing they'd predicted it, people apologising for having written that it would never happen. In English-speaking circles, it's been near silence. A couple of trade posts. Nothing that reached a general marketing feed.
Both reactions make sense locally. But neither helps if you run French pages inside a multi-market estate.
What actually happened
On 22nd July, Google launched AI Overviews and AI Mode in France, two months ahead of the 23rd September ceiling it had given French press publishers in a private letter, and with no signed remuneration agreement in place. France was the last major market without the feature. The block was legal, not technical, and it stopped holding.
The consequence for you is narrow and specific: your French organic numbers are about to move for a reason that has nothing to do with anything your team did this quarter.
The pattern from every studied market holds. Informational and top-of-funnel pages absorb the drop. Transactional, branded and local queries barely shift. Germany, structurally the closest comparison, saw click-through on the top organic result fall from 27% to 11% where an AI Overview appears. That's a split, not a collapse. But a split still lands in a quarterly report as one line going the wrong way.
Don't read the first datasets too closely
Last week, I searched for the altitude of Strasbourg. A single public figure, one line on Wikipedia, the kind of thing a featured snippet has handled cleanly for a decade. I got an AI Overview.
I also tested AI Mode on personal purchases. It kept missing whole parts of what I was asking for. Not wrong, but often beside the point.
The system hasn't found its trigger threshold yet. Anything measured this month measures a feature still settling into place, still rolling out, which is exactly what makes the French studies now circulating so confident and so premature.
The thing worth doing while it settles isn't analysis
It's a baseline.
Pull your French property's pre-22nd-July numbers now, split by query type, before the quarter closes and the comparison window fills with post-launch weeks. Search Console holds sixteen months, so this is still sitting there. It won't be interesting in October, and it won't be recoverable in the form you need.
Because the Q4 conversation is coming, and by default it will be the wrong conversation. Someone will look at a French traffic line that dipped in Q3 and reach for the explanations to hand: content, competitors, a technical regression, the agency. Separating what Google did from what your team did requires a before. Most French sites won't have one, for the same reason almost nobody in France used two years of German data: the thing wasn't here yet, so it wasn't real yet.
France spent two years watching other markets sit the exam with the answers face-up on the table. The exam started last Wednesday. The advantage was never the delay. It was what anyone did with it.
From my desk
I published a note the morning of the launch: what's confirmed, what isn't, and why the figures in circulation range from −15% to −65% without contradicting each other.
Read the AI Overviews French Launch note
The only response I got to the last issue of this newsletter was one line: "finally, an analysis with an actual action in it." I was pleased, and slightly worried. If that's what's missing, there are a lot of us commenting and not many saying what to do on Monday.
I'm also finishing a full guide on page-level AI Overview exposure. I'll send the link when it's live.
Work with me?
If you have French pages inside a multi-market portfolio and you'd rather walk into Q4 planning with the split already mapped, reply to this email. No scripted discovery call, just a look at where you actually stand.
Have a lovely day,
Alizée Baudez
This was originally sent as an issue of Beyond Organic, my newsletter on international SEO and AI search, on 30 July 2026. If you'd rather get the next one in your inbox than find it here later, subscribe here.
France gets to see the answers first
France is about to get the most mature version of AI Overviews Google has built, without the two years of trial and error other markets got to watch first. Here's the actual deployment timeline, what German data shows about which pages are exposed, and why that's a genuine advantage with an expiry date.
Hey,
Last week half my feed declared it settled. AI Overviews are coming to France, the deal is done, and a good number of people wanted you to know they'd been ready all along.
So I went back to the sources.
Here's what's actually confirmed. On 29th June, Google notified French press publishers that AI Overviews and AI Mode will launch in France by 23rd September at the latest. That part is real. Much of what came after it is not. There is no signed remuneration agreement with publishers. The APIG (a publishers association) is still asking for one. The 23rd September date is a maximum deadline in a private letter, not an announced launch. And as I write this, nothing is live on google.fr.
"Notified" doesn't mean "settled", and that gap is where people quietly plan against a date that doesn't exist.
But the timing isn't the interesting part. This is.
France doesn't get the version other markets got.
Here's the actual timeline, because it changes how you should read this.
Rough timeline of the deployment of AI Search in Google
Google unveiled the Search Generative Experience in May 2023, a labs experiment, US-only. In May 2024 it became AI Overviews and went live to everyone in the US. Within a fortnight it was telling people to put glue on pizza 🍕 and eat rocks 🪨, and Google throttled it, dropping its frequency to around 9% of queries while it added guardrails. Through late 2024 and 2025 it clawed back, expanded to a hundred-plus countries, reached nine EU markets in March 2025, and kept maturing. By early 2026 it was appearing on roughly 48% of searches.
That's two years of public trial and error. Two years of Google shipping, breaking, pulling back, and rebuilding in full view, which gave every early market a long, visible warning while the thing took shape.
France gets none of that runway. France likely goes from zero to the 2026 build in one step. And the 2026 build is a different animal from the 2024 one. When Ahrefs re-ran their study, the click-through drop on the top organic result had gone from 34.5% to 58% as the feature matured. It's more present, more confident, and better at keeping the click than anything the early markets met at launch.
There's a legal weight to it now too: a German court ruled this year that because AI Overviews produce "independent, new, and substantive statements," Google is the author of them, and liable for what they get wrong. This is a mature system arriving all at once, not an experiment easing in.
So France hasn't been through the same slow adjustment as everyone else, just later. France skipped the adjustment and inherited the outcome.
Here's the part that should change how you feel about it.
If you have a French online presence, the instinct is to read all of this as bad news arriving faster. I'd read it the other way.
France is the only major market that gets to see the answers before sitting the exam. Every English-speaking market flew blind. Germany, which got AI Overviews in March 2025, is now a fully documented case study sitting right next door in a structurally similar market. And what the German data shows is not uniform collapse. It's a split. Informational and top-of-funnel content took the hit, some parenting and health sites down 24% and beyond. Transactional, branded, and local queries barely moved. Booking and Amazon lost almost nothing.
That pattern holds in every market that's been studied. Which means it's not a forecast for France. It's closer to a map. 🗺️
You already know which of your pages are informational and which are transactional. You already have the query mix in Search Console. The work that English-speaking sites had to do reactively, in the dark, mid-drop, you get to do now, calmly, before anything moves. That is a genuine advantage, and it has an expiry date.
The uncomfortable truth under the "we've been ready" chorus is that being aware of AI Overviews and being prepared for them are not the same thing, and the second one is unevenly distributed. Knowing it's coming is not a strategy. Knowing which of your pages are exposed, and doing something about those specifically, is.
I've spent the last year mapping how this rolled out across other markets, precisely to answer the question everyone in France is about to ask at once: what actually happens to my traffic, and what do I do before it does.
From my desk
A few things I've published since the last issue, if you want to go deeper:
I'm also finishing the Pre-Flight Digital Expatriation Checklist, the practical companion to how I think about entering a new market before the first line of content goes live. I'll share the link with you once it's live.
Work with me?
If any of this maps onto your situation, the most useful next step is usually a conversation. No scripted discovery call, just a look at where you actually stand. Reply to this email if that's useful.
Have a lovely day, 😎
Alizée Baudez
This was originally sent as an issue of Beyond Organic, my newsletter on international SEO and AI search, on 9 July 2026. If you'd rather get the next one in your inbox than find it here later, subscribe here.
You're in the search results. You're not in the answers.
You can rank first for a query in one market and be completely invisible when an AI tool answers that same query in that language. Translated content quietly fails the test that written-for-market content passes, and the fix is narrower than a full overhaul.
Hey,
You can rank first in a market and be completely invisible to AI in that same market. Not ranked low, but absent.
That gap is what this issue is about.
Most international sites were built for translation. The local-language version exists because a launch plan said it should. Content was moved across languages, not written for them.
For traditional search, this is survivable. Domain authority and link signals carry a translated page further than it deserves.
For AI search, it breaks.
When a tool like ChatGPT or Claude receives a query in German about your category, it isn't assessing keyword relevance. It's looking for content it can extract: sentences that stand on their own and carry enough specificity to mean something without context. AI tools don't retrieve full pages. They retrieve passages.
Translated content fails this test. No local examples, no local vocabulary, nothing that tells an AI tool: this was written for this audience, not at them.
The distinction worth holding: translated for a market, written for a market, and built for AI visibility in a market are three different things. Most brands have only ever done the first, and most don't realise the other two exist as separate problems.
The consequence: in markets where your content was translated but not locally written, AI tools fill the gap with whoever they're more confident about. Usually a competitor who's been building consistent signals in that language for longer.
The fix isn't an overhaul. Pick the one topic per market you most need to be known for. Get it rewritten from scratch by someone who thinks in that language. The specificity difference is significant. AI citability follows.
I put this full framework into a structured checklist: 24 items across five sections, covering entity signals, content quality, technical setup, local authority, and monitoring. Each item is scored per market in a companion Google Sheet that maps where you actually stand.
The International GEO Checklist — €49, companion audit tracker included.
From my desk
This newsletter has always been a space for thinking out loud. The International GEO Checklist is the first time I've taken something from my consulting practice and turned it into an actual product — a framework I built over years of auditing international sites.
To be honest, for me it's a bigger step than it looks from the outside. There's more in the works, and I'm looking forward to sharing it with you.
And since you're here, and since you've been part of this long enough to know what I'm about, I'd like to offer you €10 off at checkout with code FIRSTLAUNCH10.
€10 off, valid until 12th May 2026.
Cheers, 🌸
Alizée BAUDEZ
Work with me
If what I write resonates with where you're navigating right now, the best way forward is a conversation to see whether it makes sense.
This was originally sent as an issue of Beyond Organic, my newsletter on international SEO and AI search, on 30 April 2026. If you'd rather get the next one in your inbox than find it here later, subscribe here.
A decade of not compromising.
My consultancy turns ten this year, and this issue is a look back at what actually held it together through Covid, a maternity leave, and a decade of algorithm updates. It started with two strangers who needed their CVs rewritten and €150 invoiced, and somehow became this.
Hi,
Today is a strange and good day.
My business turns ten this week. I've been sitting with that number for a few days, trying to figure out what to say about it. And honestly, what keeps coming back isn't a highlight reel.
It held.
Through slow quarters and strong ones. Through Covid, a maternity leave, every algorithm update, and every moment where I genuinely wasn't sure what came next. The consultancy held. And I think that's worth saying out loud, because it didn't happen on a whim.
It started in 2016 with two strangers who needed their CVs rewritten. I registered the business to invoice €150. And somehow, that was the beginning.
After two years, I had a side business generating 15% of my income on evenings and weekends, and a stubborn conviction that I wasn't going back to an office. The maths said it could work. The conviction said it had to.
What followed was a couple of years working from Bologna, Gdańsk, Vilnius, Budapest, the Canary Islands – and more – doing international SEO while actually living it. Then settling Strasbourg, one of the beating hearts of Europe.
Ten years later, I'm still here. Still independent. Still finding this work genuinely interesting.
If you want to read the full story, I've put it together at alizeebaudez.com/10th-anniversary.
I want to say thank you, specifically to you, because you're here. Whether you've been reading since the beginning, found this newsletter recently, worked with me, referred me, or simply stayed curious about what I'm building: it matters more than a standard thank-you covers.
This newsletter exists because I have things worth saying and people worth saying them to. You're one of those people.
From my desk
I'm heading into the next decade with the clearest sense of direction I've had yet. A lot of what I've been applying with clients for years (the way I think about market entry, content architecture, the decisions that happen before any technical work begins) is taking shape as something more formal. You'll be the first to know when it comes out.
If you know someone who should be reading this, I'd be glad if you passed it on.
Here's to what comes next.
Cheers, 🥂
Alizée BAUDEZ
Work with me
I work with experienced marketing teams on upstream international SEO strategy : the decisions that happen before tactics.
This was originally sent as an issue of Beyond Organic, my newsletter on international SEO and AI search, on 25 March 2026. If you'd rather get the next one in your inbox than find it here later, subscribe here.