Branded visibility flatters you. Unbranded is the number that counts
If the question already contains your brand name, the answer was never really in doubt. Most AI visibility tracking mixes these prompts in with everything else, and the result is a number that looks healthy for reasons that have little to do with whether buyers can find you.
Why branded prompts prove so little
Take a prompt like: evaluate Acme as a provider of payroll software. The model has been told who to talk about. It will look for Acme, find Acme's own pages, and describe Acme. Being named in that answer is close to guaranteed.
That is useful in one narrow sense, because it shows whether the engine describes you accurately. It says nothing about whether a buyer who has never heard of you will be pointed your way.
The split to make
Sort every prompt you track into one of two groups before you look at any results.
| Group | Example | What it shows |
|---|---|---|
| Branded | Is Acme good for payroll in Germany? | Whether you are described correctly |
| Unbranded | Best payroll software for a 200-person company in Germany | Whether you are found at all |
Comparison prompts that name you alongside competitors, such as Acme or Rival for payroll, sit in between. Keep them in a group of their own, because they only happen once a buyer already knows you exist.
How to build the unbranded set
Start from the questions buyers ask before they have a shortlist. They describe a problem or a category, often with a size or a location, and they do not mention anyone by name.
Check every draft prompt for your brand name, including product names and your domain. If any of them appear, the prompt belongs in the branded group, however natural it sounds.
Aim for enough unbranded prompts to cover your main categories and markets. Twenty well chosen ones will tell you more than a hundred that quietly include your name.
How to report it
Report the two groups separately, and lead with unbranded. A strong branded number next to a weak unbranded one is a common and important finding. It means engines know who you are and still do not think of you.
Track the unbranded number over time on a fixed set of prompts. That is the figure that moves when content and outreach start to work.
What to do on Monday
Pull your current prompt list and tag every line as branded or unbranded. Most teams find their tracking leans far more branded than they realised. Then rebuild the unbranded half until it reflects how a buyer who has never met you would actually ask.
Frequently asked questions
What counts as a branded prompt?
Any prompt that contains your brand name, including your product names and your domain. It does not matter how natural the wording is. If the engine has been told who to talk about, the prompt is branded, and being named in the answer tells you very little.
Are branded prompts worth tracking at all?
Yes, for a different reason. They show whether engines describe you accurately, which matters for reputation and for catching mistakes. Keep them separate from unbranded prompts, because mixing the two makes your visibility look better than it is.
Where do comparison prompts fit?
Prompts that name you next to competitors belong in their own group. They reflect buyers who already know you exist, so they measure how you fare within a shortlist rather than whether you reach one.
How many unbranded prompts do we need?
Enough to cover your main categories and markets, and no more than you can run consistently. Around twenty carefully chosen prompts on a fixed schedule give a clearer signal than a long list that quietly includes your brand.
See how you show up today.
Five questions, answered in minutes: an estimated visibility score, the competitor most likely taking your citations, and the fix categories that matter. No card, no commitment. Prefer to talk it through first? Book a consultation with a strategist.