Google just published an official guide on optimizing for AI Overviews and AI Mode.
We have been telling clients this for two years. Now it's officially on Google's dev blog.
Their thesis: AI search is SEO. AI features retrieve from the same index, run on the same ranking systems. Real GEO is doing the SEO work AI systems retrieve.
Google's exact phrase for what doesn't help, in scare quotes: "AEO/GEO hacks."
Specifically:
llms.txt and AI-specific text files
Chunking content for AI
Rewriting copy for LLMs
Structured data as an AI-search requirement
Inauthentic mentions in blogs and forums
What still matters is the same thing that mattered before:
Unique content with first-hand experience.
Google's example: "Why We Waived the Inspection & Saved Money: A Look Inside the Sewer Line" beats "7 Tips for First-Time Homebuyers."
Crawlable, indexable, fast pages
Real images and video where they help
What this means for ecom specifically:
We run SEO for 7-figure DTC brands. Last week I pulled AI Overview citations for one client's top 50 commercial queries. Their site got cited in 8. The other 42 went to competitors, review sites, and Reddit threads.
The 8 they got cited in share a pattern. Not the schema markup. Not the rewriting tricks.
Real content:
Product pages with founder or expert-written specs, not Shopify default descriptions
Real photos (product on a real human, in real light, not the stock studio shot)
First-party reviews with substance, not five-star one-liners
Collection pages that actually answer "best X for Y" instead of "Our [Category]"
A real About/Why-We-Built-This page the brand stands behind
Most DTC brands have zero of these.
They have AI-generated product descriptions, stock photos, scraped review widgets, and an About page their copywriter wrote in 30 minutes.
Then they wonder why AI Overviews cite their competitor instead.
Google's guide doesn't say anything we haven't been saying. It just makes it harder for the GEO industry to keep selling llms.txt audits to ecom brand owners who don't know better.
If you can't rank in Search, no LLM is going to retrieve you. That's the whole game.
Here is how we made a health focused brand €198k in a single month through SEO
No PPC. No paid social. No influencer drops.
Three things built it: topical depth, intent-mapped content and link building.
Most people in supplements and wellness assume the ceiling is content volume. Pump out 200 articles a month, hope something sticks. It doesn't work like that anymore. Volume without structure just buries you in your own thin pages.
Here's what actually moved it.
1. Topical authority
Topical authority isn't "we wrote a lot of articles." It's "this site covers the entire topic, including the subtopics most competitors skipped."
Before we wrote anything for this brand, we mapped the whole niche. Every parent topic, every supporting subtopic, every adjacent question, every product-level page that needed to exist. Then we pulled the top 10 ranking sites on the head terms and listed every page they had. The gaps in that map became the content plan.
The result is a site that owns the cluster, not a site with random posts. Internal linking ties it together so authority flows from the strong pages to the rest.
When Google sees a site that covers a topic this completely, it treats new pages on that topic differently. Indexing is faster. Ranking from new pages is faster. The site stops being one of many and becomes one of the answers.
2. Intent matching
Most content fails not because it's badly written, but because it doesn't match what the searcher actually wants.
"Best supplements for X" is a buyer searching.
Selling there works. Education there gets ignored.
"How does X work" is someone researching.
Selling there gets bounced. Education there builds trust.
"X side effects" is someone looking for safety info.
Either intent there fails unless the page is dedicated to that question.
Every keyword we target gets mapped to an intent before the brief is written. Buyer pages have a CTA above the fold. Research pages don't. Comparison pages have tables and pros and cons.
The format follows the click, not the keyword.
This is the part most agencies skip because it's slower. It also roughly doubles conversion rate on the same traffic.
3. Link building
This is the part most ecom SEO advice pretends doesn't matter. It does.
Topical depth and intent matching get you ranked on long-tail and mid-tail. Real money keywords are competitive enough that on-page alone won't get you there. You need links from sites Google trusts.
Health is YMYL territory (your money your life).
Google puts heavier weight on E-E-A-T signals here than in most niches. Links from real sites with real audiences move the needle. Links from networks and link farms either do nothing or actively hurt.
For this brand we run a steady cadence of placements every month on sites we vet manually.
No drops. No PBN networks. Direct publisher relationships, paid at the publisher's actual rate.
We got tired of platforms draining our budgets. So we built RankJet, the link building platform that connects you directly with publishers. No broker markup.
That's the part most people get wrong on cost.
Skip link building and you cap around the long-tail ceiling. For a brand in a competitive niche, that means leaving most of the addressable revenue on the table.
What I'd change
The first stretch with this brand we under-invested in cluster depth and over-rotated to link building.
Links pointing at thin content don't compound. Links pointing at a real topical hub do.
Build the content layer first. The link layer hits 2 to 3x harder when it lands on something already structurally strong.
That's the playbook. €198,361 in a single month. All organic.
@harleyf@Shopify@OpenAI Entering new era of SEO. Whoever wants to get a free checklist on how to increase the visibility on AI shopping, contact us here: https://t.co/60Z0ts08XA
BITCOIN will be $350,000 by August 25, 2024 is not a lie. It’s a prediction. It’s speculation, it’s an opinion, but it’s not a lie. It’s suckers bait, but it’s not a lie becuase any prediction about the future is not a lie. I want $350,000 BITCOIN to be a fact, and I want it to be true…. but it is only a prediction. I am confident Bitcoin will hit $350,000 sometime in 2024 but for it’s only a target, a dream, and a wish. I keep buying more Bitcoin, Ethereum, and Solana, because quite certain their prices will continue to rise. Why am I so confident? Is it because I am certain about BC, Ethereum, and Solana? My answer is “No.” What I am confident of is the incompetence of our leaders, President Biden, Treasury Secretary Yellin and Fed Chair Powell. Those three are the 3-Stooges in real life and I am certain, I am very confident about their incompetence. Take care. Buy more Gold, Silver, Bitcoin, Solana, and Ethereum. Protect your self from the 3- Stooges.