When a customer asks ChatGPT "what's the best sustainable running shoe?", the answer it gives isn't a list of ads. It's a curated recommendation, usually three to five brands, presented as if they're the definitive answer.
If your brand isn't in that list, you don't get a lower ranking. You don't exist.
That's AI visibility: whether AI assistants mention, recommend, and endorse your brand when customers ask questions about your category.
You might have heard this called GEO (Generative Engine Optimization). It's the emerging discipline of optimizing your brand's presence in AI-generated responses, the same way SEO optimizes your presence in traditional search results. The term was coined in a 2023 research paper and has since become the industry shorthand for this entire space.
But here's the thing most GEO advice gets wrong: it treats all AI systems as one channel. They're not. And understanding the differences is what separates useful strategy from guesswork.
AI visibility is already reshaping how people discover and buy products.
AI is becoming the first step in the buying journey
Search engines gave you ten blue links. You clicked through, compared, and decided. The brand's job was to rank.
AI assistants work differently. They don't send customers to your site to browse. They answer the question directly. "The best sustainable running shoes are Allbirds, On, and Veja. Here's why."
For a growing number of customers, that answer is the research. They don't click through to page two. They don't compare five tabs. They take the AI's word for it, or at least use it to narrow the field to two or three brands they then investigate further.
This matters for ecommerce brands because the point of influence has moved upstream. By the time a customer lands on your site, they may have already been told by an AI assistant that your competitor is the better choice, or that you don't exist at all.
Five AI systems, five different answers
There isn't one AI. There are several, and they work differently:
AI Search pulls live web data to answer questions in real time:
- Perplexity, an AI-native search engine. Searches the web, reads pages, synthesizes an answer with citations. Updated in real time.
- Google AI Overviews, the AI summary box that now appears above traditional Google results for many queries. Uses Google's search index.
AI Knowledge draws on training data, massive snapshots of the internet compiled months ago:
- ChatGPT, OpenAI's flagship model. Training data has a knowledge cutoff several months in the past.
- Claude (Anthropic), increasingly used for research and comparison queries.
- Gemini (Google), Google's own AI model, distinct from AI Overviews.
The crucial insight: these two channels can tell completely different stories about your brand.
A brand might be well-known in ChatGPT's training data (because it was prominent six months ago) but invisible in Perplexity (because its current web presence is weak). Or vice versa: a newer brand with strong review coverage might show up in AI Search but be completely unknown to training-data models.
Understanding which channel you're strong in, and which you're weak in, tells you exactly where to focus.
What determines whether AI mentions your brand?
This is the core question of GEO, and the answer is fundamentally different from SEO.
AI models don't rank brands the way Google ranks websites. There's no direct equivalent of Google's PageRank. AI isn't counting links. But domain authority still correlates with AI visibility, because high-authority pages are better represented in training data and rank higher in the live indices that AI Search tools draw from. The mechanism differs from SEO, and it's not the primary lever, but it's not irrelevant either. AI synthesizes information from the sources it's been trained on or can search in real time, and those sources tend to be well-established, well-linked-to pages.
The factors that matter most:
1. Third-party coverage
AI models trust independent sources. If review sites, buying guides, comparison articles, and industry publications mention your brand, AI is more likely to include you in its answers.
This is the single biggest lever. A brand that appears in Wirecutter, Good Housekeeping, or niche category review sites will show up in AI responses far more often than one that only has its own marketing content.
2. Comparison content
When customers ask "Brand X vs Brand Y", AI needs comparison content to draw from. If that content exists on your domain, on review sites, or on comparison platforms, AI can reference it. If it doesn't exist, AI either skips you or hedges with generic statements.
Brands that proactively create comparison content on their own site ("How we compare to [competitor]") give AI something concrete to work with.
3. Review signals
AI models aggregate review sentiment from across the web, not just star ratings but the actual language of reviews. Consistent positive sentiment across multiple platforms (Trustpilot, Google Reviews, Reddit, niche forums) strengthens your AI Trust score.
4. Structured data
Schema markup helps AI understand what your brand sells, where you operate, and how you relate to other entities. Product schema, Organization schema, and FAQ schema all make it easier for AI to extract accurate information about your brand.
5. Entity recognition
Does the internet "know" your brand exists as a distinct entity? Wikidata entries, Wikipedia articles, and Google Knowledge Graph presence help establish your brand as a real, notable entity, not just a keyword. This is confirmed to influence Google's own systems (including AI Overviews). Whether it directly affects training-data models like ChatGPT and Claude is less clear, but entity recognition correlates with the broader web presence that AI learns from.
This is particularly important for smaller or newer brands. Without entity recognition, AI models may not distinguish your brand name from common words.
What doesn't work
The AI visibility space is new enough that misinformation is rampant. A few things that sound plausible but don't actually help:
llms.txt files: a proposed standard for telling AI crawlers about your site. No major AI platform (OpenAI, Anthropic, Google, Perplexity) has confirmed they read it. Read more about why.
"Submitting" content to AI models: some tools claim to notify AI models about your content. Training-data models can't be pinged into updating. Live AI search tools use their own crawlers and indices. No third-party tool has direct access to push content into any AI model.
Keyword stuffing for AI: writing content "optimized for AI" by cramming in AI-friendly phrases doesn't work. AI models evaluate the overall quality, authority, and relevance of content, not keyword density.
Paying for AI placement: as of now, you cannot pay to appear in ChatGPT, Claude, Gemini, or Perplexity responses. There is no AI equivalent of Google Ads (yet). This may change, but today, AI visibility is entirely organic.
How to measure your AI visibility (and start doing GEO properly)
The first step of any GEO strategy is knowing where you stand. You could manually type queries into ChatGPT, Perplexity, Claude, and Gemini, then read every response looking for your brand name. For one query, that takes five minutes. For the dozens of queries that cover your full customer journey (discovery, comparison, and trust) it takes hours. And you'd need to repeat it for every competitor.
That's what SeenByAI automates.
SeenByAI queries real AI systems with the specific questions customers ask about your product category. It measures every mention, every comparison outcome, and every sentiment signal, then scores your visibility across two channels (AI Search and AI Knowledge) and three customer journey stages (Discover, Consider, Trust).
The result: a clear picture of where you stand and where your competitors stand, plus a prioritized Playbook of specific actions, with effort ratings and timing estimates, to close the gaps. Not just a score. A fix plan.
The opportunity is now
AI visibility is in its early days. The brands that understand it now, that measure where they stand and take specific action to improve, will have a structural advantage as AI becomes a bigger part of how customers discover and choose products.
The brands that wait, or that install a magic file and hope for the best, will wonder why their competitors keep showing up in AI responses and they don't.
SeenByAI finds where competitors are beating you in AI, and get a prioritized plan to close the gap. Get started free.
Written by Stu Miller, Founder of SeenByAI and CEO & Co-founder of Smart Insights. Stu has spent 16 years helping ecommerce businesses grow their digital marketing, and built SeenByAI after experiencing the AI visibility problem first-hand running his own Shopify store.
Related reading: