A Query Design Guide for Getting Chosen by AI Search
In GEO for EC and retail, registering product names and brand names alone is not enough.
Designing queries around the contexts in which users consult AI
—problems, use cases, comparisons, and reputation—
is the starting point for being chosen by AI.
Users don't necessarily ask AI "Should I buy ◯◯?" right away.
For example: "Recommend a toner that works for sensitive skin," "What's a good gift for Mother's Day?," "Compare popular health food brands for me," "Are the reviews for ◯◯ not so bad?"—in this way, they consult AI in the contexts of problems, use cases, comparisons, and reputation.
With Genview, we design these queries across three stages: TOFU, MOFU, and BOFU.
TOFU: The stage where the user doesn't know your brand yet
TOFU is the stage where the user hasn't yet decided on a specific product name or brand name.
At this stage, you check whether your brand makes it into the candidate set for a given problem or use case.
For example, for skincare products you would build queries starting from the user's problem or usage scenario, such as "a toner for sensitive skin," and for food gifts, "a gift recommended for Mother's Day."
If your company doesn't show up here, from the AI's perspective it may not be recognized as a "representative option for that problem or use case."
MOFU: The comparison and consideration stage
MOFU is the stage where the user is comparing multiple products or brands.
At this stage, you check whether the AI can explain the reasons your company is chosen when compared against competitors.
For example, you register category-comparison and competitor-comparison queries such as "Compare recommended toner brands for me" or "Tell me the difference between Company A and Company B."
In MOFU, it's important to look not only at whether your company is included as a comparison candidate, but also at which evaluation axes you are discussed on—price, quality, reviews, ingredients, ease of use, and so on.
BOFU: The stage just before purchase
BOFU is the stage where the user is checking their last concerns right before purchasing.
At this stage, you check whether the AI can accurately explain your brand or product.
For example, these are queries like "What are the reviews for ◯◯?," "Is ◯◯'s subscription easy to cancel?," or "Which is better, ◯◯ or △△?"
In BOFU, you check whether the AI is putting out incorrect or outdated information about reviews, effectiveness, cancellation terms, refund guarantees, competitor comparisons, and so on.
How to think about which queries to register for EC & retail
For EC and retail, we recommend including at least the following five contexts.
Context
Purpose
Query example
Problem
See if you enter the candidate set for solving the problem
What's a good product for dry skin?
Use case
See if you're recommended for the usage scenario
What's a good gift for Mother's Day?
Category
See if you're recognized within the product category
What are some recommended toner brands?
Comparison
See the reasons you're chosen over competitors
Which is better, Company A or Company B?
Reputation
See if pre-purchase concerns are answered correctly
What are the reviews for ◯◯?
It's important to register not just product categories and brand names, but the words users are actually likely to use when consulting AI.
Ready-to-use query examples
TOFU Awareness-stage query examples
Recommend a toner that's easy to use even for sensitive skin
What skincare products do you recommend for dry skin?
What product do you recommend for someone buying protein for the first time?
Recommend a health food that's easy to keep up with every day
What beauty gift do you recommend for Mother's Day?
What's a practical gift you can buy for under 5,000 yen?
Tell me how to choose products so I don't fail when shopping online
What beauty supplements do you recommend for women in their 30s?
Recommend daily necessities for someone living alone
Recommend products that are easy to use together with children
What health food subscription brands do you recommend?
Compare protein brands available online for me
Compare food brands suitable for gifting for me
Tell me the difference between Company A's and Company B's health foods
Compare the reviews of Company A and Company B
Which should I choose, Company A or Company B?
What skincare brands have prices that are easy to keep up with?
BOFU Purchase & decision-stage query examples
What are the reviews for ◯◯?
Tell me about ◯◯'s reputation
Is ◯◯ really effective?
Can ◯◯ be used even on sensitive skin?
Is ◯◯'s subscription easy to cancel?
Does ◯◯ have a refund guarantee?
Which is better, ◯◯ or △△?
Tell me who should and shouldn't buy ◯◯
Lay out the pros and cons of ◯◯
Where is the most cost-effective place to buy ◯◯?
Summary
In GEO for EC and retail, rather than looking only at product names and brand names, it's important to design queries for each context in which users consult AI.
In TOFU, the problems and use cases of users who don't know your brand yet. In MOFU, the consideration axes of users who are comparing competitors. In BOFU, the concerns and reputation checks right before purchase.
By registering queries across these three stages, you can see how your brand is recognized, recommended, and compared on AI.
Frequently asked questions
Q. Is it enough to register only queries that contain our own brand name, like "What are the reviews for ◯◯?"
Queries containing your own brand name (BOFU) are essential for checking how AI describes your company, but they aren't enough on their own. Many users considering a purchase consult AI with problems or use cases—like "a toner for sensitive skin"—while they still don't know your brand name. By also registering TOFU and MOFU queries that don't contain your brand name, you can check whether AI lists you as a candidate at the point of first contact with new users.
Q. If we sell on an e-commerce marketplace, will AI recommend us?
Listing on a marketplace contributes somewhat to AI recognition, but it doesn't guarantee a recommendation on its own. Because AI assembles a picture of your brand from multiple sources, consistency of information across your own site, marketplaces, and external media builds trust. Along with query design, it's important to organize the product information on your own site. For more detail, see our overview page on GEO for EC & retail.
Q. How can we check whether the queries we registered are producing results?
With Genview, you can continuously measure the appearance status and scores for each registered query across ChatGPT, Gemini, Perplexity, and more. You can track in numbers how AI's description changed—and whether the gap with competitors narrowed—before and after running a measure. We recommend starting with the free plan to check your current AI recognition and understand your starting point. See pricing plan details.