Branded vs Unbranded Prompts: Where AI Visibility Is Won
Short answer: branded prompts tell you how AI engines describe you after someone already knows your name. Unbranded prompts tell you whether people who do not know your name will ever hear it. For most companies, the second group is where growth is decided, and it is also where visibility is hardest to earn. Track both, and be suspicious of any score that leans on the first.
- Branded prompts are easy to “win” and often misleading as a headline metric.
- Unbranded prompts measure discovery, which is the real competition.
- Competitor-named prompts sit in between and are worth separate tracking.
- The two groups are fixed by different things: your own site and facts for branded, independent coverage for unbranded.
Two kinds of question, two different jobs
Take a company that sells project management software. Compare these questions:
- “Is Acme PM good for small agencies?”
- “What’s the best project management tool for a small agency?”
The first buyer already knows Acme. The question is about trust and fit. The engine will usually find Acme’s site, a few reviews and some discussion, and the answer will mention Acme because the question demands it.
The second buyer does not know Acme exists. The engine chooses which brands to name, and Acme either makes the list or does not. A company can be well described in the first conversation and absent from the second, and if it is absent from the second, the number of new buyers it reaches shrinks.
That is the distinction. Branded prompts are about reputation. Unbranded prompts are about reach.
Why branded numbers flatter you
If you track a prompt that contains your brand name, the engine will nearly always mention you. It would be an odd answer otherwise. So a prompt set heavy with branded queries will show a mention rate that looks strong and rises very little with effort.
Teams get caught by this in three ways:
The dashboard looks healthy. A 90% mention rate sounds great until you learn that 80% of the prompts say your name.
Progress looks flat. You publish, earn coverage and win threads, yet the branded rate does not move because it was already near the ceiling. The real change is in unbranded prompts that were never tracked.
Competitive reads are skewed. Share of voice on branded prompts mostly reflects who the prompts are named after. We cover why in AI share of voice.
None of this makes branded tracking pointless. It means a branded rate should not be your headline.
What branded prompts are good for
Used properly, branded prompts are a reputation and accuracy check. They show you:
- What the engine says about your pricing, features and limits, and whether it is current. Outdated or wrong claims travel far.
- How you are framed. “Affordable but limited” and “enterprise-grade” are different positions.
- Which competitors are named next to you. If the engine always pairs you with a rival you do not consider a peer, that tells you how you are perceived.
- Which sources define you. Often it is a review site or an old article, not your homepage.
Run these monthly. They are cheap, and a wrong answer here is usually fixable by updating pages and profiles. The practical steps are in how to get your brand mentioned in ChatGPT answers.
Why unbranded prompts are the real battlefield
1. They are where new buyers appear
Most buyers who are not yet your customers will not type your brand. They describe a problem or a category. An unbranded answer typically offers a short list of names. Being on that list is the AI-era equivalent of ranking on page one for a category keyword. Being off it is invisibility, and you will not see the loss in any branded metric.
2. The list is short
Search gave you ten blue links. A generated answer might name three to six brands. Competition for those slots is tighter than competition for a position on a results page, and the winners are often the same few names for a given prompt.
3. Your own site matters less
For branded prompts, your site is a primary source. For unbranded recommendation prompts, engines tend to look for independent evidence, since “this company says it is the best” is weak. That shifts the work from your own pages to what others say about you: roundups, reviews and discussions.
Reddit is a clear example. Threads titled like buyers’ questions are retrievable and are cited for recommendation-style prompts. In ReddGrow’s study of 994 Reddit threads that AI engines cited, the strongest signal was how well the thread’s title and opening matched the prompt, and roundups and guides were among the most often heavily cited formats. Citation share varies by engine and over time, as we explain in why ChatGPT cites Reddit, so treat it as one source among several.
4. They respond to category-level work
Unbranded visibility moves when your category presence grows: being included in comparison articles, accumulating genuine reviews, being recommended in community threads, publishing the clearest explanation of a problem. That is slower than editing a product page, and it compounds.
Competitor-named prompts: the third group
Prompts like “alternatives to [Competitor]” and “[Competitor] vs [You]” do not fit neatly in either bucket. A name appears, but the buyer is still choosing. They are some of the highest-intent prompts you can track, because the buyer is already unhappy or undecided.
Track these separately. The sources cited tell you where comparisons are made, and your position in the answer tells you whether you are seen as a credible alternative. Our post on AI competitive analysis for brand mentions shows how to read them.
A simple framework
| Prompt type | Example | What it measures | Main lever |
|---|---|---|---|
| Unbranded, problem | ”Why do patients miss appointments?” | Early discovery | Content that answers the question, community presence |
| Unbranded, category | ”Best scheduling software for dental clinics” | Category inclusion | Independent roundups, reviews, discussions |
| Competitor-named | ”Alternatives to [Competitor]“ | Share of a rival’s demand | Comparison coverage, accurate positioning |
| Branded | ”Is [You] worth it?” | Reputation and accuracy | Your own pages, profiles, reviews |
When you report, show the four rows separately. The unbranded rows are the story for growth, and the branded row is the story for risk.
How to rebalance a prompt set
If your current list is mostly branded, here is a straightforward fix.
- Count. Tag each tracked prompt as unbranded, competitor-named or branded.
- Rebuild toward 70/30. Add unbranded prompts from sales calls, support tickets and community threads. See how to choose prompts to track.
- Keep the old data separate. Do not blend history across a changed prompt set. Start a new baseline for the unbranded group.
- Report the groups separately. A single blended rate will move for reasons you cannot see.
- Set targets on the unbranded group. That is where you would like the line to rise.
Reading the pattern in your results
Once you track both, the combination is diagnostic:
- High branded, low unbranded. The engine knows you and does not recommend you. You need independent coverage and presence in the category conversation.
- Low branded, low unbranded. Awareness problem. Start with fundamentals: crawler access, accurate profiles, a clear description, then third-party mentions.
- High unbranded, low branded. Unusual. You may be recommended without a good official source to describe you. Strengthen your own pages.
- High both. Maintain, watch competitors, and check by engine, since strength in one can hide weakness in another.
What to do next
Most of the work in raising unbranded visibility is unglamorous and slow: earn honest coverage in the places engines already cite, answer the questions buyers ask in public, and publish the clearest explanations in your category. Tooling helps with the tracking. ReddGrow tracks your prompts across engines and shows which Reddit threads are behind the citations, which is useful for the unbranded side. You can get a first read with the free Am I on AI check.
Takeaway
If you remember one thing: a branded prompt asks “what do you say about me?” and an unbranded prompt asks “who should I consider?” Only the second brings new buyers. Build your prompt set and your reporting so that second question carries the weight.
Frequently Asked Questions
- What is the difference between branded and unbranded prompts?
- A branded prompt names a company, such as 'Is Acme good?' or 'Acme vs Beta.' An unbranded prompt names none, such as 'best invoicing software for freelancers.' Unbranded prompts are where buyers who do not know you yet discover options, so they measure discovery. Branded prompts measure how you are described once you are known.
- Why is my brand visible for branded prompts but not unbranded ones?
- Because the engine knows you exist and can describe you on request, but does not consider you among the best answers for the category. That usually points to thin independent coverage: few comparison articles, reviews or community discussions that name you alongside your category. Fixing it means earning that third-party presence.
- Should I stop tracking branded prompts?
- No. Branded prompts reveal inaccurate descriptions, outdated pricing, negative framing and which competitors engines pair you with. They are cheap to monitor and useful for reputation. Just do not let them dominate your prompt set or your headline metric.
- Are competitor-branded prompts worth tracking?
- Yes. Prompts like 'alternatives to [Competitor]' and '[Competitor] vs [You]' sit close to a buying decision, and the sources cited for them show where comparison happens. They count as branded in the sense that a name appears, but they behave like discovery for a rival's audience.
- What ratio of branded to unbranded prompts should I use?
- A reasonable starting point is around 70% unbranded and 30% branded, including competitor-named prompts. Lesser-known brands should lean even more unbranded, since branded prompts will often return little. Adjust once you see which group your decisions depend on.