The Ultimate Customer Research Survey For GEO

Consider this an open kimono.

To date, we’ve not shared the survey below publicly. However, my team got tired of people in the GEO echo-chamber ignoring inalienable facts, so we are releasing this beast into the wild.

If there’s an ulterior motive, it’s for sanity to prevail over slop.

The Problem This Survey Exists to Solve

Most brands approach GEO backwards.

They jump straight to optimization tactics, like schema markup, content production, Reddit activations and PR, without asking the questions that should come first.

Specifically, these are:

Which AI tools are our buyers and prospects using?
How do they query LLMs?
When in the buying journey do they consult AI?

Why this matters

As has been noted ad nauseam, LLMs are not interchangeable. Muck Rack’s ongoing What Is AI Reading? research has now analyzed more than 25 million links cited across ChatGPT, Claude, and Gemini. The data clearly shows the top cited websites shift meaningfully across models, which means you need to pick the right LLMs to optimize for. That process begins with understanding client behavior.

So let’s move the horse to its rightful spot, in front of the proverbial cart. The survey framework below gives you empirical data on the tools your buyers prefer, when and how they use them, and how AI impacts purchase behavior. GEO strategy and tactics flow from that foundation.

How to Use This Information

We’re purposely sharing a heavy version of our comprehensive framework. It is organized across seven sections and 30 questions. A few important points:

No mandate will require that you ask every question here. Think of our survey as a menu rather than a mandatory checklist. In our experience, a focused study of 15 to 18 well-chosen questions will outperform a bigger 30-question instrument every time, because respondent fatigue is real.
Read through all the sections first, identify the hypotheses that matter most for your specific strategic situation and then select the questions that test those hypotheses most directly.
The sections do not need to be fielded in the order presented here. Indeed, any individual section could be lifted out or used independently.
The [product category] placeholder used throughout should be replaced with your specific category before you go out to the field for research.
If your category is broad, consider whether it should be split into subcategories to produce cleaner behavioral and attitudinal data.

Section 1: Tools Your Customers Use

As we said, before you can make any meaningful decisions about where to invest in GEO, you need to know which AI tools your buyers are using. This sounds obvious, but it is the step many brands skip entirely, defaulting instead to assumptions based on aggregate data.

The gap between Question 2 and Question 3 will reveal actionable insight. A customer might use ChatGPT most frequently overall but switch to Perplexity specifically when researching purchases in your category because they trust its sourcing more. If you only ask Question 2, you risk optimizing for the wrong platform.

Q1. Which of the following AI tools have you used in the past 90 days? Select all that apply.

ChatGPT
Perplexity
Google AI Overviews / Google AI mode
Claude
Gemini
Microsoft Copilot
Grok
Other (please specify)
I have not used any AI tools in the past 90 days

Q2. Which of the following AI tools do you use most frequently overall? (Single select from the list above via a dropdown menu)

Q3. Which AI tool do you use most frequently when you are researching a purchase in [product category]? (Single select from a dropdown menu, with “I do not use AI tools for this purpose” as an option)

Q4. How long have you been using AI tools regularly?

Less than 6 months
6 to 12 months
1 to 2 years
More than 2 years
More than 3 years

Section 2: How Customer Use LLMs

Knowing which tool a customer uses tells you where to focus. Knowing how they use it tells you what kind of content and authority signals you need to build. A customer asking a quick lookup question requires a very different GEO response than one conducting deep comparative research before a significant purchase.

Q5. When you use an AI tool to research a purchase in [product category], what are you typically trying to accomplish? Select all that apply.

Discover a product that solves a problem I have
Compare two or more specific solutions to a problem
Get a recommendation based on my needs or budget
Understand technical specifications or features of potential products and solutions I might use
Find the best available price or current deals on specific solutions
Read a summary of reviews or opinions from other buyers
Other (please specify)

Q6. When you use an AI tool to research [product category], how would you describe a typical session?

I ask one question and act on the answer
I ask one question and then follow up with a few more related questions
I have a multi-part, back-and-forth conversation with the tool
It varies considerably depending on what I am trying to find out

Q7. How much do you trust AI-generated product recommendations compared to the following sources? Rate each on a scale of 1 (trust much less than AI) to 5 (trust much more than AI).

A review from a major publication like Wirecutter, CNET, or The Verge
A review from a friend or family member
Customer reviews on Amazon
A recommendation from a social media influencer or creator
The brand’s own website
Counsel or marketing from a retailer like Best Buy or Target

Section 3: AI Use Across the Purchase Funnel

This is the section most brands find surprising when the data comes back. The prevailing assumption is that AI tools matter primarily at the top of the funnel, around awareness and discovery.

Our experience, backed by this interesting data set on AI use around GLP1 purchases, is that this assumption is wrong, and that AI is typically used throughout the buyer’s journey, including into post-purchase. There are meaningful commercial implications at each stage.

Q8. Thinking about the last time you purchased a product in [product category], did you use an AI tool at any of the following stages? Select all that apply.

Before I knew what I wanted — I used AI to understand what type of product I needed
Once I knew the category — I used AI to identify which brands or models to consider
When comparing options — I used AI to evaluate specific products against each other
Right before buying — I used AI as a final check on my decision
During the purchase — I used AI to find the best price or deal
After I bought — I used AI to learn how to set up or get the most out of the product
After I bought — I used AI to troubleshoot a problem
After I bought — I used AI to decide whether to keep, return, or exchange the product
After I bought — I used AI to research accessories or complementary products
I did not use AI at any stage of this purchase

Q9. At which single stage did an AI tool have the most influence on your final decision? (Single select from the same list above, excluding all “After I bought” answers and “I did not use AI”)

Q10. When you most recently used AI to research [product category], what triggered that session?

A product I owned broke or stopped working
I saw an ad or piece of content that made me curious
Someone recommended a product to me and I wanted to learn more
I had a problem and started researching from scratch
I was comparison shopping after already identifying candidates
Other (please specify)

Q11. After purchasing a product in [product category], which of the following have you used an AI tool for? Select all that apply.

Getting setup or onboarding help
Learning advanced features or tips
Troubleshooting a problem or error
Deciding whether to return or exchange the product
Writing or informing a review I posted
Finding accessories or add-on products
Comparing it against a product I am now considering as a replacement or upgrade
I have not used AI tools after a purchase

Q12. How often do you use an AI tool at more than one stage of the same purchase journey?

Almost every time I make a purchase in this category
Occasionally
Rarely
I typically only use AI at one stage per purchase
I do not use AI during purchases

Section 4: Query Behavior

This section feeds directly into Prompt Universe mapping, which is the exercise of identifying the 100 to 200 specific queries your buyers are most likely to ask LLMs. Knowing this allows you to build content and authority signals around the right prompts.

Q13. When you ask an AI tool about [product category], which of the following best describes how you typically phrase your questions? Select all that apply.

I ask for a specific product recommendation (e.g., “what is the best wireless earbud under $100”)
I describe a problem I want to solve and ask for suggestions (e.g., “I work out every day and keep losing earbuds, what should I get”)
I ask it to compare specific products I already have in mind (e.g., “compare the Sony WF-1000XM5 and the Bose QuietComfort Earbuds”)
I ask about price ranges or value (e.g., “what is a good [product] for around $150”)
I ask technical questions about features or compatibility
I ask about deals or sales (“are there places offering discounts on Bose Quiet Comfort earbuds now”)
I ask if there are problems associated with the product (“have there been common criticisms or problem with Bose Quiet Comfort Earbuds”)

Q14. When you use an AI tool to research [product category], whose language do you tend to use?

I use technical terms and industry language (e.g., “ANC,” “THD,” “IP68 rating”)
I describe things in everyday language (e.g., “earbuds that block out noise on a plane”)
It depends on how much I already know about what I’m looking for
I’m not sure

Q15. When you start an AI research session for [product category], are you typically already aware of specific brands or models?

Yes, I usually have two or more brands or models in mind already
Sometimes. I usually know one brand but am looking for alternatives
No, I usually start with a problem or need and let the AI surface options

Q16. When comparing products using an AI tool, which of the following do you most commonly ask it to compare? Select all that apply.

Two specific products you already identified
A specific product against “the best alternatives”
Products within a defined price range
A brand you know against brands you don’t
The current model of something against an older version you already own

Q17. Which of the following have you asked an AI tool about a specific brand in [product category]? Select all that apply.

Whether the brand is reputable or trustworthy
Known problems or complaints about the brand’s products
How the brand compares to its main competitors
Whether the brand offers good value relative to its price
The brand’s customer service or warranty reputation
I have not asked AI about specific brands

Q18. When an AI research session leads you toward a product you hadn’t considered before, what most often caused that shift? Select all that apply.

The AI surfaced a brand I wasn’t aware of
The AI reframed my problem in a way that changed what I was looking for
The AI cited a source that changed my thinking
The AI identified a feature I hadn’t considered but immediately wanted
This hasn’t happened to me

Q19. Have you ever purchased a product specifically because an AI tool recommended it?

Yes, I purchased the exact product the AI recommended
Yes, I purchased a product from a brand the AI mentioned, though not the exact model recommended
I used the AI recommendation as one input among several but it influenced my final decision
No, I have used AI tools for research but have not made a purchase based on a recommendation
No, I do not use AI tools for purchase research

Section 5: Source Trust and Verification Behavior

This section tells you what happens after AI makes a recommendation, which is just as strategically important as the recommendation itself. Understanding how customers evaluate what AI says, and what they do after an LLM makes a recommendation, are critical for your GEO program success.

Q20. If an AI tool recommends a product to you, what do you typically do next? Select all that apply.

I click through to a source the AI cited or linked to
I search for the product on Google to read more
I go directly to Amazon to check reviews and pricing
I go directly to the brand’s website
I ask the AI follow-up questions
I ask someone I know for their opinion
I go to a retailer like Best Buy to see it in person
I act on the recommendation without doing further research

Q21. Which types of sources do you trust most when an AI tool cites them in a recommendation? Rank your top three.

Major tech or consumer publications (Wirecutter, CNET, The Verge, PCMag)
General news publications (NYT, WSJ, Washington Post)
Brand or manufacturer websites
Amazon customer reviews
Reddit discussions
YouTube reviews
Specialist or niche publications in the product category
Influencer or creator content

Q22. Has an AI tool ever given you product information that turned out to be inaccurate, bad, or outdated?

Yes, and it affected my trust in AI recommendations significantly
Yes, but it did not meaningfully affect my trust in AI recommendations
Not that I am aware of
I always verify AI recommendations so accuracy of the initial answer matters less to me

Section 6: General Trust

We’ve found that trust in AI tools correlates with a number of downstream purchase behaviors, so we suggest asking this question.

Q23. How much do you trust AI tools generally, given what you know, or have heard, about the companies that build them?

I trust them a great deal
I trust them somewhat
I am neutral
I am somewhat skeptical
I do not trust them

Section 7: Demographics and Regional Data

Cross tabulation is where the party’s at. These questions are the variables you cross-tabulate against the data above. On their own they are unremarkable. Combined with the behavioral and attitudinal data from the preceding sections, they are where actionable intelligence emerges.

A few of the most useful cuts to run:

High-value buyers versus the general population. Customers who purchase multiple times a year often behave very differently from occasional buyers and their AI preferences may be different.
Age against AI tool preference. Simple enough, but insightful.

Q24. What is your age?

18 to 29
30 to 39
40 to 49
50 to 59
60 and older

Q25. What is your annual household income?

Under $35,000
$35,000 to $74,999
$75,000 to $124,999
$125,000 to $199,999
$200,000 or more
Prefer not to say

Q26. What is your gender?

Man
Woman
Non-binary or gender non-conforming
Prefer not to say

Q27. Which state do you currently live in? (Dropdown — used to assign Census region on the back end: Northeast, Midwest, South, West)

Q28. Which of the following best describes the area where you live?

Large city or urban area (population over 1 million)
Mid-size city (population 100,000 to 1 million)
Small city or large town (population 10,000 to 100,000)
Suburban area outside a city
Rural area or small town (population under 10,000)

Q29. How frequently do you purchase products in [product category]?

This would be my first purchase in this category
Once every few years
About once a year
Multiple times a year

Q30. How would you describe your familiarity with [product category]?

I am new to this category
I have some experience
I consider myself knowledgeable
I am an enthusiast or expert

That’s it. You might have more to add.

Keep in mind, this is just a tool. Pick your questions, get into the field, and let the data tell you where to go. Your cash will thank you.