Your brand can show up in AI search and still be left out of the recommendation. So what is missing?
When ChatGPT, Gemini, Perplexity, or Google AI Overviews answers a question, the sources supporting the response are not always the same brands mentioned in the answer itself. A company may appear in the citations while another company is named as the one to consider.
That distinction is becoming increasingly important as businesses invest in SEO, content, press releases, and AI visibility. A citation tells you that an AI system found your information relevant enough to reference. A recommendation goes further. It means the system has enough confidence to put your brand directly into the answer.
So the question is no longer simply whether AI can find your brand. It is whether AI has enough evidence to choose it.
Everybody in this space has been selling you on getting cited. Turns out getting cited is the participation trophy. When researchers looked at what happens after a brand shows up as a source, the brand got left out of the actual recommendation 69% of the time that's cited and ignored. Hey, let me show you the difference and how you actually get recommended by ChatGPT. Stick around. A robust tool for enhancing your digital footprint. A business owner, marketer, and know who loves that customer. You guys. Ready to make headlines? Here's what that number is actually measuring because it matters. When an AI answers a question, two separate things happen. It pulls sources, those little footnotes, the links, the according to material, and then it writes the answer. And in the answer, it names things, brands, products, company. It says, go with this one. Those are not the same list. You can be in the footnotes and completely absence from the sentence that actually sends someone to a checkout page. Roughly two out of three times, that's exactly what happens. Which means a whole lot of people are celebrating a metric that really doesn't move revenue. It's like being thanked in the credits of a movie, you're not in. It's very nice, call your mom. But nobody's buying a ticket because of it. So why the gap? Because sourcing and endorsing are two different jobs. Sourcing is a relevance question. Is this page related to what was asked and can I point at it? That bar is low. You wrote about the topic, you're indexed, you're crawlable. Hey, congratulations, you're a source. That's the participation trophy I was talking about. Endorsing is a confidence question. The model is about to put its name on a recommendation and the thing about a recommendation is that it's a claim. This is the good one. That's a much higher bar and it needs more than one page saying so. Think about how you would handle it. Somebody asked for a plumber. You've read one guy's website saying he's the best plumber in three counties. Are you passing that name along? Probably not. Maybe you've heard it from three people who don't know each other. Models work on the same instinct. One source is an input. Multiple independent sources say the same thing is a pattern, and patterns are what get named in the answer. Okay. So what actually tips a citation into a recommendation? Three things. None of them are secrets. One, corroboration. The same facts about you showing up on sources that aren't yours and aren't connected to each other. Your website saying it is table stakes. Six independent outlets saying it is evidence. Two, specificity. Models are bad at repeating mush like industry leading solutions provider gives it nothing to handle to a person. Serves 340 clinics across 12 states and founded in 2009, gives it something concrete to say out loud. Vague copy is unrecommendable. I know that's not a word, but you get what I mean. Three, recency. A recommendation is a bet on the present tense. Information from four years ago carries a risk. Did they close? Did they move? Did they get acquired? Fresh material lowers that risk and lower risk gets named. Corroboration, specificity, recency. If you're cited but never recommended, you're almost certainly missing the first one. Now, this is where I'd normally sell you something. I'm going to, but stick with me because the mechanics are the actual point. Corroboration is a distribution problem. You can't write your way into it. 10 blog posts on your own domain are still one voice repeating itself. A press release solves it structurally. One release syndicated across hundreds of independent outlets, different domains, different publishers, no relationship to each other, all carrying the same verified facts about your company on the same date. That is the exact pattern a model reads as corroboration. And when you build it with the SCAN framework, structure, clear focus, authoritative quotes, numerical specificity, you're hitting cooperation and specificity and recency in one move. That's not me being clever. That's just what the format already does. Hey, we've had releases picked up in an AI overview inside of twelve hours. Before you buy anything from anyone, including me, run this first. It takes just five minutes. Probably not even that actually. Prompt one. What are the best, your category, companies in your city? Are you in the answer or just the sources? Prompt two. Tell me about your brand name. Does it know real specific facts or is it reading your homepage back to you? And then prompt three, who would you recommend for the exact service you sell? That's the one that matters. That's the money question. Cited on all three, recommended on none, that's a cooperation problem. And a cooperation problem is fixable. Run the three prompts on your own brand. Seriously, do it before you close this tab. You'll learn more in five minutes than from another forty minute AI SEO webinar. If you want a deeper version of this, hey, I wrote a free book, how to get Google's AI to feature your brand. The link to the book is in the description. No cost, no catch. And if you run the test, don't like what you see? Grab a free consultation with me. We'll look at your actual results together and find where the corroboration is missing. The link to my calendar is in the description. Citation was never the finish line. Getting chosen is. Like, follow, subscribe, and I'll see you in the headlines.
The News Guy approaches this as a practical AI visibility problem. Getting cited is useful, but citation alone does not necessarily translate into recommendations, leads, or customers. In fact, research discussed in the video found that after a brand appeared as a source, it was left out of the actual recommendation 69% of the time.
That creates a meaningful gap between being visible and being chosen. Closing that gap requires more than publishing another page. It requires stronger corroboration, greater specificity, and information that is current enough to support a recommendation.
Key Takeaways
- A brand can appear in AI citations without being named in the actual recommendation.
- Sourcing and recommending are different decisions that require different levels of confidence.
- Corroboration gives AI systems multiple independent sources supporting the same facts about a brand.
- Specific, verifiable information gives AI something concrete to use instead of vague marketing claims.
- Recency helps establish that the information being used still reflects the current business.
- Press releases can help corroborate information by distributing consistent, newsworthy details across multiple outlets.
- The three-prompt test can help determine whether your brand is being cited, recommended, or overlooked.
If you’re building AI visibility around your brand, it also helps to understand how GEO differs from traditional SEO and why appearing in search is only one part of becoming a source that AI systems can understand and use.
What Does “Cited But Not Recommended” Mean?
Being cited and being recommended may sound like the same thing, but they represent two different outcomes. When an AI system cites a source, it is indicating that the source contains information relevant to the question. That information may help explain a company, product, service, location, or other subject.
A recommendation requires something more. If someone asks, “Who are the best plumbers in my area?” an AI system might find your website and use it as a source without naming your company among its recommendations. Your website may contain useful information, but the system may not have enough supporting evidence to confidently present your business as one of the best choices.
The difference is similar to researching a business versus recommending one to a friend. Finding information is relatively easy. Putting your reputation behind a recommendation requires greater confidence. That is why appearing in an AI citation shouldn’t be treated as the finish line.
Ready to get Started?
Why Can a Brand Be Cited Without Being Recommended?
AI search has to make more than one decision when constructing an answer. First, it needs to determine which information is relevant enough to use. A website that contains useful information about a particular company, service, or topic may therefore become one of the sources behind the response.
Then the system has to determine what belongs in the actual answer. That is where brands can disappear from view. A company might provide useful information about a particular service and therefore become a source, while another company has stronger evidence across multiple independent sources supporting its reputation, location, specialization, or other characteristics relevant to the question.
The first company gets cited. The second may get recommended. For a business, that difference matters because the recommendation is much closer to the point where a potential customer decides what to investigate, contact, or buy.
What Makes a Brand More Likely to Be Recommended?
Three factors provide a useful framework for understanding the difference between citation and recommendation: corroboration, specificity, and recency.
Corroboration: More Than One Voice
Your website can tell the world that your company is excellent. That information has value, but it still comes directly from you. Corroboration happens when independent sources provide consistent information about the same business, product, service, or achievement.
Think about asking a friend for a plumber recommendation. One plumber’s website might describe the company as the best in the area. That tells you what the plumber thinks. Three unrelated customers telling you the same plumber is reliable provides a different kind of evidence because the information is being independently repeated.
For businesses, press release distribution can contribute to this pattern when the same verified facts about a company or announcement appear across multiple independent news and media outlets. The objective is not to manufacture mentions. The goal is to make important business facts available beyond the company’s own website.
Specificity: Give AI Something Concrete to Say
AI systems have little use for vague promotional language. “Industry-leading solutions.” “World-class service.” “Best-in-class technology.” These statements may sound impressive, but they provide little concrete information that an AI system can confidently use when comparing businesses.
Specific information is different. A statement such as “serves 340 clinics across 12 states and was founded in 2009” gives an AI system identifiable facts it can incorporate into an answer. This is one reason structured press release writing matters. The SCAN Framework emphasizes Structure, Clear Focus, Authoritative Quotes, and Numerical Specificity, giving important information a clearer form for retrieval and citation.
Vague copy gives an AI system very little substance to work with when it has to explain why one brand should be considered over another.
Recency: Recommendations Are About the Present
A recommendation is inherently current. If an AI system finds information about a business from four years ago, it must consider whether that information still reflects the company today. Has the business moved? Has its service changed? Was the company acquired? Is the product still available?
Recent, clearly dated information gives AI systems a stronger reference point for understanding the current state of a business. That makes recency particularly important for companies whose products, locations, services, or market positions change over time. Corroboration, specificity, and recency work together to provide a stronger foundation for moving from being referenced to being considered for recommendation.
Why Corroboration Is a Distribution Problem
This is where many AI visibility strategies run into a limitation. You cannot create meaningful independent corroboration simply by publishing ten articles on your own website. Ten articles from the same domain are still one source.
The information becomes more useful when important facts about the business appear consistently across independent publications, organizations, and other relevant sources. That is one reason press releases remain useful in an AI-driven search environment.
A well-structured announcement can distribute the same verified information across multiple outlets, creating more opportunities for independent sources to reference the company, announcement, product, location, or milestone. The goal is not to create noise around the brand. It is to establish a consistent information pattern that extends beyond the company’s own domain.
How Press Releases Can Support AI Recommendations
A press release provides something that ordinary website content often does not: a newsworthy reason for multiple external sources to discuss the same business at the same time. A product launch, funding announcement, expansion, partnership, milestone, or major company development can give publishers a specific story to reference.
When the release is structured clearly, the important facts are easier to identify. When it is distributed broadly, those facts have more opportunities to appear across independent domains. When the information is recent, it also provides a current reference point for the brand.
This is where structure and distribution work together. The SCAN Framework helps make an announcement clearer and more specific, while press release distribution creates opportunities for those facts to be corroborated outside the company’s own website. Neither approach guarantees an AI recommendation. Together, however, they can address several of the conditions that make brand information easier for AI systems to understand, evaluate, and potentially use in a recommendation.
The 3-Prompt Test: Is Your Brand Being Cited or Chosen?
Before investing in another AI visibility tactic, test your current position. Open ChatGPT, Gemini, or another AI search platform and ask three questions.
1. “What are the best [your category] companies in [your city]?”
Look at the actual answer, not just the sources underneath it. Is your company named, or does your website simply appear among the citations?
2. “Tell me about [your brand name].”
Look at the facts the system provides. Does it know specific information about your company, or does it simply repeat generic language from your homepage?
3. “Who would you recommend for [the exact service you sell]?”
This is the most important test. You are no longer asking whether AI knows who you are. You are asking whether it would actually choose you.
If your company is cited across these prompts but rarely recommended, the issue may not be basic visibility. It may lack corroboration, specificity, or current evidence to support the recommendation.
What to Do If You’re Cited but Not Recommended
Start with the evidence rather than immediately creating more content. Review the information AI systems already associate with your brand. Look for vague claims, outdated information, inconsistent descriptions, and important facts that appear only on your own website.
Then identify what you actually want AI systems to understand about the business. That could be a particular service, location, specialization, product, milestone, or area of expertise. From there, create a focused announcement around something genuinely newsworthy and structure it so the important facts are clear, specific, attributable, and current.
A press release can then distribute those facts beyond your own domain, while supporting content can reinforce the same information on your website and other relevant properties. The objective is not simply to create more mentions. It is to create a clearer and more credible pattern around the information you want AI systems to understand.
Building a Longer-Lived Recommendation Signal
Don’t treat AI visibility as a one-time ranking event. Businesses change. Products evolve. Services expand. Locations open and close. New milestones replace old ones.
That means the information supporting a recommendation needs to remain useful over time. Regularly updating important business information, publishing genuinely newsworthy developments, and maintaining consistent facts across authoritative sources can help keep the brand’s digital footprint current.
A single announcement can establish a useful signal, but continued corroboration and fresh information can make that signal more durable. The goal is not simply to appear in an AI answer once. It is to build enough relevant, specific, and current information that your brand has a stronger foundation for being considered when similar questions are asked in the future.
Citation Is Not the Finish Line
AI visibility has often been measured by a simple question: Did the brand appear in the citations? That is useful, but it is incomplete. A brand can be visible in the sources and still lose the recommendation.
The more meaningful question is whether AI systems have enough confidence to understand what the business does, connect it with specific facts, and identify it when someone asks which company they should consider. That is why corroboration, specificity, and recency matter.
A structured press release can make important facts easier to retrieve. Distribution can place those facts across independent sources. Supporting content can reinforce the same information and keep the brand’s digital footprint current. Together, these approaches create a stronger foundation for AI visibility.
Citation gets your brand into the conversation. Recommendation puts it in the answer.
Moving From Cited to Chosen
Being cited by an AI system is a useful visibility signal, but it does not necessarily mean the system will recommend your brand. Citation answers whether your information is relevant enough to reference, while recommendation requires greater confidence that the brand is the right choice.
You can support that confidence with corroboration, specificity, and recency. Independent sources can reinforce the same facts, specific information gives AI systems something concrete to repeat, and current information helps establish that those facts still reflect the business today.
To evaluate where your brand currently stands, start with the three-prompt test and see whether your company appears in the answer or only in the sources. You can also learn more about building stronger brand signals through the Google Entity Stack and Cloud Site Stack approaches.
For a broader framework on making announcements easier for AI systems to understand, the SCAN Framework for AI-Friendly Press Releases provides a practical approach to structuring important facts, authoritative quotes, and numerical details.
If you want a practical starting point, download the free guide, How To Get Google’s AI to Feature Your Brand, for additional strategies for improving how your brand is discovered and referenced across AI-powered search.
You can also schedule a free 15-minute consultation to review your current AI visibility, discuss your goals, and identify where stronger corroboration may help move your brand from being cited to being considered.
Frequently Asked Questions
What does “cited but not recommended” mean in AI search?
“Cited but not recommended” means an AI system uses a brand or website as a source but does not name that brand in its actual recommendation. The information is relevant enough to reference, but the system may not have enough confidence or supporting evidence to recommend the business directly.
Why does my brand show up in AI citations but not the answer?
A citation indicates that AI found your information relevant to the question. Being recommended requires additional confidence that your brand is an appropriate choice, which can be influenced by factors such as independent corroboration, specific business information, and recency.
How can I increase my chances of being recommended by AI?
Start by making your business information specific, consistent, and current across relevant sources. Independent coverage, structured announcements, and other authoritative references can provide additional context and corroboration, although no strategy guarantees an AI recommendation.
What is corroboration in AI search?
Corroboration is the presence of consistent information about a brand across multiple independent sources. When the same key facts are supported beyond the company’s own website, AI systems have more external evidence to understand the brand.
Do press releases still matter for AI visibility?
Press releases can contribute to AI visibility by distributing timely, structured information about a company, product, service, or announcement across multiple media and news sources. Their value depends on the announcement’s quality and newsworthiness, its structure, and the resulting coverage.
How do I know if my brand is recommended or just cited?
Ask an AI search platform a direct recommendation question, such as “Who would you recommend for [your service] in [your location]?” If your company appears in the supporting sources but not in the actual answer, you are being cited without being recommended in that response.
Can I be recommended by AI without using press releases?
Yes. Press releases are one potential source of external corroboration, but AI recommendations can draw on many types of information and sources. The broader objective is to establish accurate, specific, current, and independently supported information about your brand.
Why does specificity matter for AI recommendations?
Specific information gives AI systems concrete facts they can incorporate into an answer. Details such as locations, dates, numbers, services, products, and verifiable achievements are generally more useful than broad promotional claims.
Does recency affect AI recommendations?
It can. Businesses and their offerings change over time, so older information may be less reliable for a current recommendation. Recent, clearly dated information can provide stronger evidence about what a business currently offers or represents.