A marketing manager told me: "We checked all the queries we registered in Genview — we're showing up in all of them."
When I looked at their registered queries, this is what I saw:
- [Brand name] overview
- [Brand name] pricing
- [Brand name] reviews
Every single one contained their brand name. I asked: "What happens when someone who's never heard of your company asks AI for help?"
There was a pause. They scrolled through the list again. Not one query without their brand name existed.
This isn't an unusual case. Most companies starting with Genview are in exactly this state. "We show up in AI" actually means "we show up when someone already knows our name." For someone who doesn't know the brand yet, the company isn't even in the running. This is GEO's most common blind spot.
Why this mistake happens
The cause is simple: people are designing GEO queries the same way they think about SEO keywords.
In SEO, users search short phrases like "GEO tool" or "GEO comparison." Checking your own brand name in rankings feels natural. But people talk to AI differently.
- How do I check whether my company is being recommended by AI?
- What should I look for when choosing a GEO tool?
- What AI search optimization tools are available as of 2026?
There's conversational context — not just keywords. And most users don't know your brand name yet when they start asking. That means checking only branded queries leaves the most critical moment completely invisible: the moment a potential customer asks AI for help without knowing you exist.
The problem isn't the number of queries. It's which customers, and which conversations, you're actually looking at.
Query design has three axes
To solve this, I've come to think about query design across three axes.
Axis ①
Funnel design
Segment by
consideration stage
×
Axis ②
RAG trigger design
Mix in queries that
prompt AI web search
×
Axis ③
Named / unnamed design
Do you appear
without your brand name?
Axis ①: Funnel design — segment by consideration stage
How a user asks AI depends entirely on where they are in their decision process. ToFu, MoFu, and BoFu give a useful framework.
ToFu: Becoming aware of the problem
The user isn't looking for a specific tool yet. They're asking questions like "What is AI search optimization?" and "Why isn't my company showing up in AI responses?" Your brand name doesn't need to appear here. The goal is to confirm that AI is correctly explaining the problem space your company operates in.
- What is AI search optimization?
- What is GEO?
- Why isn't my company showing up in ChatGPT?
- What's the difference between SEO and GEO?
MoFu: Looking for solutions
The user has recognized the problem and is now looking for concrete approaches and tools. The key here is whether the "GEO tool" category appears in AI responses.
- How do I check whether my company is being recommended by AI?
- What is a GEO tool?
- What should I look for when choosing a GEO tool?
- What should I check first for AI search optimization?
BoFu: Comparing and evaluating options
The user is now actively comparing specific tools. It's not just about appearing — it's about whether your company is described accurately, with the right context and strengths.
- What are the best GEO tools to use?
- What AI search optimization tools are worth using as of 2026?
- What is Genview?
- How is Genview different from other tools?
The manager from the opening had only registered BoFu branded queries. Without ToFu and MoFu, the users just starting to recognize the problem — and those actively searching for solutions — become completely invisible.
Axis ②: RAG trigger design — mix in queries that prompt AI web search
Not every query triggers an AI to run an external web search (Grounding). General conceptual questions are often answered from trained knowledge alone. Queries asking for recent or specific information are more likely to trigger a web search.
As an example Genview launched in April 2026. It's barely in AI training data yet. For queries answered from internal knowledge only, Genview simply won't appear. That's why it's important to intentionally mix in RAG-trigger queries — ones that include words like "latest," "as of 2026," "official information," or "compare."
- Internal knowledge queries: "What is GEO?" / "What's the difference between SEO and GEO?" → Check AI's general awareness of the category
- RAG trigger queries: "What GEO tools are worth using as of 2026?" / "What are the latest AI search optimization tools?" → Check whether web-published information is being picked up
Axis ③: Named / unnamed design — do you appear without your brand name?
This is the axis that connects directly back to the opening story.
Unnamed queries contain no brand name — questions like "What GEO tools do you recommend?" or "What AI search tools work best for e-commerce?" Most users don't know your brand name when they start asking. Whether you appear in these conversations is what I consider the most meaningful indicator of GEO performance.
Named queries include your brand — "What is Genview?" "What can Genview do?" The question here isn't whether you appear, but whether you're described correctly. If named queries return inaccurate descriptions, the fix is content — your official site, FAQ, feature pages, and external placements.
Named queries only show you how you appear to people who already know you. GEO is about how you appear to people who don't.
How to think about the URLs you connect to queries
There's one more thing that matters when registering queries: whether the URL you connect to actually contains an answer to that query.
For example, if you link a feature page to the query "How do I check whether my company is being recommended by AI?", that page should contain a sentence like this:
"Genview lets you see how your brand is being introduced across multiple AI platforms — including ChatGPT, Gemini, Claude, Perplexity, and Grok."
Simply linking a URL isn't enough. The design order should be: query → the answer AI is looking for → the URL that contains that answer. Thinking in this order makes pages more likely to be cited when RAG retrieves them.
① Query
The question
a user asks AI
→
② The answer AI needs
Write a sentence that
directly answers the query
→
③ Connect the URL
Point to the page
that contains the answer
A practical query design example
Here's how Genview itself sets up queries — nine queries with linked URLs, covering all three axes from problem awareness through to comparison and evaluation.
ToFu
MoFu
BoFu
Rather than only registering queries that include your brand name from the start, including queries that reflect how users who don't yet know you ask AI for help lets you check whether you're appearing in unnamed conversations.
* The number of queries you can register and manage varies by plan. Registering 9 queries is available on the Pro plan. See the pricing page for details.
Common mistakes to avoid
Only registering named queries
The case from the opening. You can only see how you appear to people who already know you. Make sure unnamed queries are always included.
Skipping ToFu and MoFu
BoFu-only queries — "best tools," "comparison" — don't let you diagnose why you're not appearing. Is the problem category not being recognized? Is the solution category missing you? ToFu and MoFu are how you find out.
Not including RAG trigger queries
For newer services, internal-knowledge queries often return no results. Include at least one or two queries with words like "latest," "as of 2026," or "based on official information" to check whether web-published content is being retrieved.
Linking URLs without updating the content
Registering a query means nothing if the linked page doesn't contain an answer to that query. Design in order: query → answer → URL.
Summary
- Most "we show up in AI" situations are actually "we show up when someone already knows our name" — unnamed user conversations are invisible
- The problem isn't the number of queries — it's which customers and which conversations you're actually monitoring
- Funnel design: segment queries across ToFu, MoFu, and BoFu
- RAG trigger design: newer services especially need RAG-trigger queries mixed in — internal knowledge alone won't surface them
- Named / unnamed design: appearing in unnamed queries is the most meaningful measure of GEO performance
- The URL you connect to a query needs to contain a clear answer to that query
Related term: For funnel stages, see ToFu / MoFu / BoFu.
Related term: For how AI retrieves external information, see RAG (Retrieval-Augmented Generation).
Related term: For AI real-time retrieval, see Grounding.
Related column: For how AI selects which pages to read, see Stop Making 1,000 Pages. Make the 10 AI Actually Reads.
While writing this, I kept wondering whether Genview could measure "whether RAG is firing" as part of the product itself. Right now, the same query may or may not trigger a web search depending on the day, the platform, and factors we can't control. If we could score "how likely is this query to trigger RAG," it would make query design significantly more precise. We're not there yet — but it's on my list.