Blog/AI Visibility
AI VISIBILITY

Branded vs Unbranded Questions: Which AI Result Tells You More?

Two rounded search bars of dark smoked glass floating one above the other in a deep blue void, the upper one dim with a small tag chip at its left end, the lower one glowing warm amber along its length

Key Takeaways

  • Know which kind of question you are asking. A branded question has your name in it. An unbranded question has only the need, the way a buyer who has never heard of you would ask.
  • Read unbranded results for whether you get found. They are the only questions that show whether an assistant brings you up when nobody mentioned you, which is where new clients come from.
  • Read branded results for whether you are described right. An assistant usually finds you when your name is in the question, so the useful finding is a wrong price, an old name or a mixed-up city.
  • Do not let a branded score reassure you. Being named five times out of five when you typed your own name five times measures almost nothing about discovery.
  • Weight your question set toward unbranded. Keep a few branded questions to check accuracy, and give the rest of the set to the questions a stranger would ask.

The first thing most coaches do after hearing about AI search is type their own name into ChatGPT. It usually finds them, says something reasonable, and they close the tab feeling fine.

That is a branded question, and it is the easy one. The question that decides whether new clients find you is unbranded: the buyer's need with no name in it. Branded vs unbranded AI questions measure two different things, and unbranded results tell you more about being found.

Below is what counts as each kind, why the same coach can score high on one and low on the other, what each result can and cannot tell you, and how to mix the two in a set of questions.

Branded vs unbranded AI questions: which tells you more?

Unbranded questions tell you more about whether AI recommends you, because they are the only ones where the assistant has to bring you up on its own. Branded questions tell you whether what it says about you is right.

The difference comes from how the answers are built. When ChatGPT searches, it reads web pages and shows links to its sources, and OpenAI's help page on searching the web with ChatGPT says results can be incomplete, outdated or incorrect. With your name in the question, the search goes straight to pages about you. Without it, the assistant has to find you among everyone who meets the need.

Google's AI features work in a similar way. Its guide to AI features and your website says AI Overviews and AI Mode may use a "query fan-out" technique, running several related searches to build one answer. An unbranded question fans out into searches about the need, and your pages have to answer those searches to be picked up.

What counts as a branded or an unbranded question

A branded question contains your business name or your own name. An unbranded question contains only the need, the situation and maybe a city, written the way a buyer who has never heard of you would type it.

  • Branded: "Is [Name] Coaching any good?", "What does [Name] Coaching charge?", "[Name] Coaching reviews".
  • Unbranded: "Best executive coach in Denver for a new CEO", "How to find an executive coach for a founder".

The people asking are different too. A branded question usually comes from someone who already heard your name, from a referral, a talk or a post. An unbranded question comes from someone who has not, which is the buyer you are trying to reach. What happens when the branded question comes back blank is its own case, covered in what happens when AI does not know your business name.

Why the same coach can score high on one and low on the other

The same coach can score high on branded questions and low on unbranded ones because the name does the work in the first set. Put the two results next to each other and the gap is the finding.

ILLUSTRATIVEONE MADE-UP DENVER EXECUTIVE COACH, TEN QUESTIONS

The same coach, five questions with the name in them and five without

Branded: the name is in the question

  1. "Who is [Name] Coaching?" Named
  2. "Is [Name] Coaching any good?" Named
  3. "What does [Name] Coaching charge?" Named, old price
  4. "[Name] Coaching reviews" Named
  5. "[Name] Coaching or a bigger firm?" Named

Named in 5 of 5. One wrong fact found.

Unbranded: only the need is in the question

  1. "Best executive coach in Denver for a new CEO" Not named
  2. "Executive coach for a first-time VP of sales" Named
  3. "How to find an executive coach for a founder" Not named
  4. "Executive coaching for nonprofit leaders" Not named
  5. "Is executive coaching worth it for a small company?" Not named

Named in 1 of 5. Four questions where a new client meets someone else.

Added together, 6 of 10 says little. Split the questions and the branded half found one wrong price, while the unbranded half found four questions where a new client meets someone else.

answerhalo.com
Illustrative, one made-up coach. Branded: "Who is [Name] Coaching?" named; "Is [Name] Coaching any good?" named; "What does [Name] Coaching charge?" named with an old price; "[Name] Coaching reviews" named; "[Name] Coaching or a bigger firm?" named. Five of five, one wrong fact. Unbranded: "Best executive coach in Denver for a new CEO" not named; "Executive coach for a first-time VP of sales" named; "How to find an executive coach for a founder" not named; "Executive coaching for nonprofit leaders" not named; "Is executive coaching worth it for a small company?" not named. One of five.

In this illustrative example the coach is named in 5 of 5 branded questions and 1 of 5 unbranded ones. Added together that is 6 of 10, which hides both findings. Split, each half says something specific.

The branded half found one thing to fix: an old price on a page the assistant read. The unbranded half found four buyer questions where someone else gets named, and the one question that did name the coach, about a first-time VP of sales, shows which kind of page is working.

What each kind of result can and cannot tell you

Each kind of result answers one question well and another badly. The table sets branded and unbranded results side by side, our reading of what each is good for.

Branded vs unbranded AI questions compared, our reading.
  Branded question Unbranded question
What is in it Your business name or your own name The need, the situation, maybe a city
Who usually asks it Someone who already heard your name Someone who has never heard of you
How often you are named Usually, if anything about you is online Only if your pages answer the need
What a good result tells you The facts about you are right Strangers with the problem can find you
What a bad result tells you A wrong price, name or city is out there Someone else is named for your buyer's question
The fix it points to Find and correct the page with the wrong fact Write the page that answers that need
Share of a 25-question set About 2 or 3 The other 22 or 23

Measuring across many questions, not one, is how the research does it too. The Princeton-led GEO study that coined the term generative engine optimization built a benchmark of 10,000 queries to measure how visible a source is in AI answers. One question, branded or not, is an anecdote.

How to mix the two in a set of questions

Mix the two by giving branded questions a small, fixed share and the rest to unbranded ones. Two or three branded questions are enough to check your facts; discovery needs many more, because buyers phrase the need in many ways.

  1. Write two or three branded questions: who you are, what you charge, and whether you are any good.
  2. Write the rest unbranded: the need, the situation and the city, in the buyer's words, not yours.
  3. Score them apart. Report branded and unbranded results as two numbers, never one total.

How to split the unbranded share across different kinds of buyer question is laid out in how to pick the 25 questions that matter, and which of them are worth the most effort is covered in which AI questions are worth winning.

The damaging admission: if most of your clients come by referral, branded questions matter more to you than this post suggests, because nearly every buyer arrives already knowing your name. In that case a wrong fact in a branded answer is the bigger risk, and unbranded discovery is a growth question you can take on later.

To see what ChatGPT says when buyers in your niche ask, and whether you appear, start with the free check. It asks ChatGPT 25 real buyer questions from your niche at $0 and emails a plain-English readout: your score, who got named in your place, and the one fix to start with. It runs as soon as you ask for it, and the report lands within the hour between 7am and 5pm Central, or by 10am the next morning outside those hours.

Questions we hear the most

What is the difference between branded and unbranded AI questions?

A branded question names your business, such as "Is [Name] Coaching any good?". An unbranded question names only the need, such as "best executive coach in Denver for a new CEO". The first tests what AI says about you; the second tests whether it brings you up.

Which tells me more, branded or unbranded results?

Unbranded results tell you more about whether new clients can find you, because nobody gave the assistant your name. Branded results tell you whether what it says about you is accurate. You need both, but most of your questions should be unbranded.

Why does ChatGPT know me by name but not recommend me?

Because the two questions ask different things. With your name in the question it can search for you directly. Without it, it has to connect the need to your business from what your pages and profiles say, and that link is often missing.

Should I worry if AI gets a fact about me wrong in a branded question?

Yes, because a buyer asking by name is often close to deciding. A wrong price, old name or wrong city at that point can cost the booking. Find the page still saying the old fact and fix it there.

How many branded questions should be in a set of 25?

Two or three is usually enough. They check that the facts are right, and that does not take many questions. The rest of the set should be unbranded questions a stranger with your buyer's problem would ask.

Can I write unbranded questions myself?

Yes. Write the questions a buyer asks before they know any names: the need, the situation, the city, the budget. Avoid your own phrases and service names, because a buyer who has never seen your site will not use them.

How do I see which unbranded questions name me?

Ask ChatGPT, Perplexity and Google the questions your buyers ask, without your name, and note who is named in each answer. The AnswerHalo free check asks ChatGPT 25 real buyer questions from your niche at $0 and shows who got named in your place.

What does AnswerHalo cost after the free check?

The full audit is $500 once and runs 125 checks across ChatGPT, Claude, Gemini, Perplexity and Google AI Overviews, landing by this time tomorrow. Ongoing watching is $399 a month, or $300 a month added with the audit.

SEE WHERE YOU STAND

Your buyers are already asking. Find out what AI tells them.

The free check asks ChatGPT 25 real buyer questions about your niche and sends you the report within the hour, at $0. If it shows you are already getting named everywhere, we will say so in plain words.

Get my free check →