The GEO 90-day playbook for local e-commerce: be visible when AI searches for you
Summarize this article with AI
Your shop isn’t in the AI response. You don’t know it yet.
A client calls me on a Tuesday. He runs a chain of 12 organic clothing stores. He invested $8,000 in local content. Photos. Blog posts. Google Business listings. Result? Nothing. His online sales are flat. His customers tell him: « We searched for you on ChatGPT, you weren’t mentioned. »
This isn’t an isolated case. According to an Uberall study (April 2026), 68% of local brands are completely invisible in AI engine recommendations. Meanwhile, $750 billion in consumer spending is already shifting to AI search. 60% of searches end without a single click to a website.
The problem isn’t your content. It’s your absence.
I’m Stéphane. I’ve been observing this with my clients for 18 months. SEO alone isn’t enough anymore. You need GEO – Generative Engine Optimization.
This isn’t a trend. It’s the new playing field. And you have 90 days to claim your position.
Week 1-4: Become the source of truth
The first thing AI looks at? The consistency of your core data. Name, address, phone, hours, services. One gap between Google, Facebook, your site, and a local directory, and the AI considers you less reliable. It filters you out.
The mechanism. Models like ChatGPT or Gemini learn from multiple sources. If they diverge, the model weights your entity lower. That’s what we call the source of truth.
With a client – an organic tea shop in Lyon – I found 7 different versions of his hours across the web. Result: zero AI citations for « best organic tea Lyon. » In 2 weeks, we standardized all directories, his site, Google Business Profile, and added a LocalBusiness schema. Within 30 days, he appeared in 4 local AI responses.
Immediate action: audit your NAP across 50 sources. Use a tool like Yext or BrightLocal (or do it by hand). Fix everything. Add specific service attributes. Think like an AI: every inconsistency is a distrust signal.
Result I’ve seen with my clients: +120% citations on average after standardizing core data.
Week 5-8: Engineer conversational context
Classic SEO demands keywords. GEO demands answers to questions. Users no longer type « running shoes Paris. » They tell their assistant: « Where can I buy running shoes with good advice in the 11th? »
The AI will synthesize an answer from sources it deems authoritative and useful. If your site doesn’t contain content that answers exactly that question, in a clear format, you’ll be invisible. Even if you rank first on Google.
The mechanism. This is what I call context engineering. You must map real questions (from your customer service, local forums, Google reviews) and produce content that answers directly: structured FAQ pages, local glossaries, practical guides.
A client – a children’s bookstore in Bordeaux – created 15 pages: « top 5 books for kids by age. » Each page answered one precise question with a clear question-answer format. In 8 weeks, he was cited by ChatGPT in 11% of responses related to kids’ books in Bordeaux. Zero percent before.
Action: list your 30 most frequent customer questions. For each, write a short answer (50-100 words) and a long answer (300 words) with clear structure (lists, tables). Publish them in a local « Frequently Asked Questions » section.
Week 9-12: Orchestrate social proof and authority
Once your data is clean and your context well-written, you need to prove you’re a trustworthy source. AI looks at trust signals: fresh customer reviews, mentions in local media, backlinks from authoritative sites.
The mechanism. Orchestration – the third pillar of GEO. You coordinate three types of signals:
- Reviews: 15 reviews per month per location (with responses).
- Press citations: get your shop mentioned in articles from BBC, Le Parisien, local blogs.
- Inbound links: not random, but from pages discussing your specialty.
A client – a wine merchant in Lille – received 18 citations in articles about natural wine in 6 weeks. Result: his product listing pages became Gemini’s preferred source for natural wine recommendations in the Nord region. Traffic +340% in 3 months.
The order is critical: first source of truth, then context, then authority. Reverse it and the AI won’t trust you.
Action: launch a local PR campaign. Pitch your expertise. Most importantly, respond to every customer review – that’s a powerful signal.
The counterintuitive truth: GEO is not SEO
People often tell me: « AI uses Google, so SEO alone is enough. » Wrong. AI doesn’t rank pages. It answers a question. If your site ranks first on Google for « shoes Lyon, » but the AI finds a clearer answer elsewhere (a local blog post, a Wikipedia page, a customer review), it will use that source.
I saw a site – 45% of its Google traffic – become invisible in AI responses overnight. Why? Its content was optimized for keywords, not for questions.
The real shift is here: you must optimize for recommendations, not rankings.
The good news? The same trust signals (NAP, reviews, authority) serve both. But the priority changes. GEO borrows the semantic cluster concept, adapted for AI: each page must be a standalone answer block, structured for retrieval.
How to apply this playbook to your business
You don’t have a dedicated team. You’re a shopkeeper, you already manage your site, inventory, customers. This playbook fits into 4 blocks.
- Week 1-4: Audit and standardize – your core data across all platforms. Use a spreadsheet. Fix one by one.
- Week 5-8: Produce question-answer content – not generic articles. Precise answers. Use a structured FAQ format.
- Week 9-12: Amplify trust signals – solicit reviews, press citations, local backlinks. Show you’re a reference.
- Keep going – GEO isn’t a one-time project. It’s a system. Refresh your questions every month, update your content, monitor new citations.
One tip: start with one location. Master the process at a single store before scaling. I saw a client lose 3 months trying to do everything at once. Result: zero impact.
Want a concrete number? A wine merchant in Lille followed this playbook strictly. In 90 days, he went from 0 AI citations to 47 citations in local recommendations. Organic traffic +820%. No ads.
Your turn now
AI engines are changing the rules. Your direct competitors? Maybe they don’t know yet. But in 6 months, those who started this playbook will already be cited, already referenced in responses.
I build these systems. I don’t sell packages. I construct the architecture that makes your local e-commerce visible when AI searches for you. The first call is a live audit. We look together at where you stand.
One question haunts me: how many of your customers are you losing because AI answers without citing you?
Free GEO audit – Are you visible in AI responses?
I offer a live 45-minute audit. We analyze your local e-commerce, check your presence in ChatGPT, Gemini, and Perplexity. No slides, no pitch. We look together at where the leverage is. Book your slot.
Book a strategic call — 45 minFrequently Asked Questions
What’s the difference between SEO and GEO?
SEO optimizes for ranking in SERPs. GEO optimizes to be cited as a source of truth in AI-generated responses. The goal shifts: you’re no longer trying to attract a click, but to be recommended as the best answer.
How long until I see GEO results for my local business?
Within 30 days if your core data is consistent and your content answers precise questions. First significant results arrive around 90 days, when AI starts citing you regularly. With my clients, +150% citations on average in 3 months.
Do I need to rebuild my entire site for GEO?
No. You don’t need a full redesign. The essentials: clean your NAP data (name, address, phone), add structured FAQ pages, and produce content answering real customer questions. The rest can stay as is.
Are Google reviews important for GEO?
Absolutely critical. AI models use reviews as trust signals and contextual content. The more recent reviews and responses you have, the more you’re perceived as an active, reliable source. Target 15 reviews per month per location.
Can I do GEO without paid tools?
Yes. Audit your data manually on the 10 main platforms, extract frequent questions from customer conversations, solicit reviews via email. Tools speed things up, but the method stays manual if you have time. I recommend starting without tools to understand the mechanism.
