Google Warns CMOs: SEO Tools Don’t Have Access to Its Internal Metrics – GEO Is SEO
Summarize this article with AI
What Google Just Told CMOs. And What You Need to Do About It.
I’m going to tell you something SEO agencies hate hearing.
The tools you use to measure your ranking on generative AI… they’re not worth much.
Not because they’re bad.
Because Google just reminded everyone: these tools have zero access to its internal metrics. Zero.
« Google does not evaluate third-party SEO tools or vendors directly, and they have no access to our internal metrics. »
— Brendon Kraham, VP Search & Commerce Global Ads Solutions, Think with Google
The article published on Google’s marketing platform is aimed at CMOs. The message is crystal clear. AI visibility dashboards are not connected to Google’s real signals. They infer, estimate, score. But they don’t see.
I look at 15 sites every week. Many show me red curves. « My AI score dropped 30 points! » Yet their sales are stable, organic traffic is climbing.
If a tool can’t see Google’s internal signals, why trust it with your decisions?
A Client Who Almost Broke Everything Because of a Tool
A client calls me Tuesday morning. E-commerce, 4,200 organic sessions per month, 800 products. He says: « I lost 60% of presence on Generative AI, I need to rebuild everything. » He shows me a well-known tool displaying a score that dropped from 47 to 19 in two months. Panic.
I ask him to show me his Search Console and Merchant Center data. There, I see something different. Impressions from AI features are stable, slightly up. Clicks from those same features: +8%. And revenue from organic traffic up 14%.
His tool was lying to him. Or rather, it was extrapolating from public data without understanding Google’s internal signals.
We stopped looking at that tool. We refocused effort on page structure, entities, tailored content, and Google indicators. I applied the DOSE framework (Demand, Optimization, Structure, Engagement) that I teach with Guillaume Attias at BMO Academy. We rebuilt 47 semantic clusters around queries actually generating clicks.
8 months later, overall organic traffic had surged +137%. And clicks from AI Overview even more. Without ever looking at a tool score.
This isn’t an isolated case.
Why Your SEO Tools Have Access to Nothing – And Never Will
Google owns its internal rankings, quality scores, thresholds for AI Overviews. No API distributes them. Third-party tools can only scrape search results or simulate queries. They infer presence.
But they don’t know if your page was shown in an AI Overview for 1,000 queries or 10. They don’t know the real click-through rate, or whether the answer was actually read.
« Google does not evaluate third-party SEO tools or vendors directly, and they have no access to our internal metrics. »
When Google says that, it’s not saying tools are useless. It’s saying they don’t see the full picture. Just a noisy trailer.
When I audit a site, I never ask for a tool score. I ask for Search Console reports, Merchant Center, and Google Analytics 4. That’s where I see real data. Clicks, impressions, transactions.
You can have an AI score of 90 on a tool and zero real clicks. And vice versa.
It’s brutal, but it’s liberating.
The 3 Google Reports I Use Every Day (And That Tools Miss)
First report: Search Console, Performance tab, filtered by « Appearances in AI features ». There, you see impressions, real clicks, CTR. These are concrete data. I observe with my clients that CTR on AI results is often lower, but impression volume can be massive. Your diagnosis relies on the trend.
Second report: Google Merchant Center, for e-commerce. Product listings in AI Overview are visible under « Search network » if you enable the filter. You’ll see how many products appear, clicks, and even assisted conversions via conversion tracking. A client in Southeast Asia discovered that 23% of their product clicks came from AI Overviews, which their visibility tool never flagged.
Third report: Google Analytics 4. Even without perfect segmentation, you cross-reference « organic » source with conversions. What matters is that traffic converts. Tools telling you « ready for AI » are secondary.
With these three reports, I diagnosed 945 optimizable pages in one afternoon. Without opening a single third-party platform.
GEO = SEO. Period. Here’s How the DOSE Framework Applies.
Google’s been saying it since last May: AEO and GEO are SEO. And I’ll add: good SEO.
The DOSE framework, which I apply with my clients and teach via Guillaume Attias at BMO Academy, works well with AI Search.
- Demand: identify your customers’ real questions. Not tool keywords, but conversational long-tail queries visible in Search Console. On a 320-product site, I found 47 AI queries with high volume, but only 1.2% CTR. We created dedicated answers.
- Optimization: create content that answers clearly, structured with lists, tables, FAQs, and entity data (Schema Product, FAQ, HowTo). That’s what Google extracts to generate an answer. An industrial client structured their product sheets as Q&A; their AI appearances jumped +210% in 4 months.
- Structure: organize your site into semantic clusters. Link pages together with relevant connections, build a readable entity graph. I’ve deployed over 1,300 clusters since 2016, and it works for AI too. An e-commerce site with 2,200 pages, restructured into 14 clusters, saw 39% of its AI queries generate a click.
- Engagement: clicks, time on page, bounce rate influence whether Google AI picks your page. Stop chasing backlinks. Work on expérience. A site with dated design gained 17% more AI clicks after a simple product page redesign.
GEO isn’t a new discipline. It’s just SEO applied to today’s Google ecosystem.
The Third-Party Dashboard Trap (And How I Got 47 Sites Out of It)
In Q1 2025, I audited 47 e-commerce sites. 43 of them tracked an AI visibility score with a tool. In 39 cases, that score didn’t reflect their real traffic. Worse, 12 clients had made costly decisions based on those scores: redesigns, page deletions, stopped SEO work. None had a real problem.
In my live audit, I never talk about tools. I show traffic, pages, real positions from Search Console. The client sees their data, not a dashboard. When we roll out the action plan, we base it on these native reports and the DOSE framework.
Result: of those 43 sites, 38 saw organic traffic grow in the following 6 months, without spending a dime extra on ads.
The most revealing part? 4 sites had an AI visibility score dropping according to their tool, while their revenue from organic traffic climbed 20% to 40%. Google’s indicators told the truth.
What You Can Do Tomorrow Morning, Without Any SEO Tool
Open Search Console. Filter by « AI features » over the last 3 months. Note the trend in clicks and impressions. If the curve is climbing, you’re fine, even if your tool score drops.
Go to Merchant Center. Look at product listings under « Search network ». How many AI clicks are your products generating? That’s where your visibility actually plays out.
Find AI queries with clicks in Search Console. Check which pages rank. Do they answer directly? Do they have structured data? If not, add clear content, FAQ or HowTo tags, and strengthen internal linking from your pillar pages.
Repeat this analysis in 30 days. No tool. No dashboard.
I have a client who followed this routine. He went from 12 AI clicks per month to 347 in 3 months. No secret: just reading signals correctly.
What if your favorite SEO tool has been lying to you for months?
Live SEO Audit: I’ll Show You Your Real Performance, Not a Tool Score
For 60 minutes, I dig through your Search Console, Merchant Center, and GA4 data with you. We spot blockers, find high-converting AI queries, and I deliver a custom DOSE action plan. No third-party tool will touch your data, but you’ll walk away with a clear roadmap.
Book a strategic call — 45 minFrequently Asked Questions
Why does Google say SEO tools don’t have access to its internal metrics?
No Google API publishes ranking scores, AI Overview impression rates, or answer selection criteria. Tools scrape or simulate; their reliability is limited. Google repeats this so CMOs don’t treat third-party indicators as gospel.
How do I measure my visibility on generative AI without paying for a tool?
I look at native Google reports: Search Console for AI feature impressions and clicks, Merchant Center for product listings, and Google Analytics 4 to link to conversions. Data comes straight from Google.
Is GEO really just standard SEO?
Yes, Google confirmed it. The same principles — structured content, topical authority, schema markup, user expérience — decide whether you appear in AI Overviews. The DOSE framework works for it too.
Should I stop using third-party SEO tools?
No, they’re still useful for competitive analysis, tracking traditional keywords, and technical audits. For AI Search, keep a critical eye on their scores and always cross-check with Search Console.
How do I build content that appears in AI Overviews?
I answer specific questions with clear lists, tables, and definitions. I use FAQ, HowTo, and Product schema tags. I structure pages with short paragraphs and subheadings. I link content together. That builds a coherent entity graph. It’s the semantic cluster principle.
