What is GEO? A plain guide for Shopify and D2C founders

Ask ChatGPT “what’s the best sunscreen for oily skin under ₹500?” and you won’t get ten blue links. You get a short answer that names three or four products and explains why. Whoever is in that answer gets considered. Whoever isn’t, doesn’t.

Generative engine optimisation, or GEO, is the work of getting your brand into those answers. This guide explains how AI assistants choose brands and what you can realistically do about it.

How an AI assistant decides which brands to name

There is no single ranking algorithm to reverse-engineer, but the inputs are well understood:

• What the model learned in training. Brands that are written about often and consistently, on your own site, in reviews, in articles and on marketplaces, are more likely to be “known”.
• What it finds when it searches. Many assistants now search the web while answering. If their crawlers can read your pages, your product details can be pulled into the answer.
• How clearly your pages describe the product. Structured data (schema) for products, FAQs and reviews makes it easy for a machine to understand what you sell, for whom and at what price.

GEO is not the same as SEO, but it builds on it

Good SEO still helps: fast, crawlable pages with clear content are what both Google and AI crawlers want. The difference is the outcome you measure. SEO measures your position in a list of links. GEO measures whether you are named at all, where in the answer, and how you are described.

That’s why a brand can rank on page one of Google and still be missing from ChatGPT’s answer to the same question.

Three things to do this month

1. Make sure AI crawlers aren’t blocked

Check your robots.txt for rules that block GPTBot, ClaudeBot, Google-Extended or PerplexityBot. Some themes and security apps add these blocks without telling you. We wrote a five-minute guide to checking it.

2. Answer the questions shoppers ask, on your own pages

If people ask “is this suitable for sensitive skin?”, say so clearly on the product page, and add it as an FAQ with FAQPage schema. AI answers borrow language from pages that answer the question directly.

3. Measure before and after

Pick 10 to 15 questions your customers really ask and check what each AI says today. Change something, then check again. Without a baseline you can’t tell what worked. See how to choose those questions.

What GEO can’t do

No tool can guarantee that an AI will recommend you. Models change, answers vary from one run to the next, and a better-known competitor will often be named first. What you can do is remove the obstacles, give AI clear material to work with, and track whether you are gaining ground. That is exactly what Signal AI is built for.