If ChatGPT recommends your competitors but rarely mentions your company, the problem is usually bigger than your website. AI systems may be finding stronger evidence across the web that your competitors belong in the answer.
That evidence can come from editorial coverage, reviews, comparison articles, industry publications, expert commentary, customer discussions, YouTube and other third-party sources that repeatedly associate a company with a particular category, product or problem.
For a growth-stage technology company, this creates a new marketing problem.
You can rank well on Google. You can have strong brand awareness. You can spend heavily on paid media. You can even have a successful PR program.
And ChatGPT can still recommend three competitors ahead of you when a prospective customer asks:
“What are the best companies for X?”
“What is the best alternative to X?”
“Which platform should I use for X?”
“What companies should I consider if I need X?”
Increasingly, those answers matter.
OpenAI’s documentation on ChatGPT Search explains that ChatGPT can search the web and provide citations to relevant sources. OpenAI also says search results are ranked using multiple factors intended to surface relevant and reliable information.
Google is moving in the same direction through AI Overviews and AI Mode. Google now provides website owners with specific guidance for visibility within its generative AI search experiences and, as of 2026, dedicated generative AI performance reporting in Search Console.
The question for CMOs is becoming fairly simple:
When your buyers ask AI about your category, does your company make the answer?
If your competitors consistently appear and you don’t, there are seven places I would look first.
1. Your competitors have stronger third-party authority
Start here.
Your website can say you’re the leader in your category.
That doesn’t necessarily give an AI system enough independent evidence to reach the same conclusion.
Third-party sources matter.
Muck Rack’s May 2026 What Is AI Reading? study analyzed more than 25 million links cited by ChatGPT, Claude and Gemini across 17 industries.
The results are difficult for communications teams to ignore:
- 84% of citations came from earned media
- 27% of cited sources were journalism
- 0.3% came from paid or advertorial content
Those patterns have also been remarkably consistent across three editions of the research.
For a growth-stage CMO, the implication is important.
What other people say about your company can contribute to the information environment AI systems use when answering questions about your market.
Imagine two competing enterprise technology companies.
Company A says on its website that it is an innovative leader in the category.
Company B says essentially the same thing, but it has also been reviewed by relevant technology publications, included in industry comparisons, quoted in articles about the category and discussed by independent experts.
Company B has created a much larger independent evidence trail.
That can become a competitive advantage in AI search.
2. You have press coverage, but not the right press coverage
This is where things get more interesting for PR teams.
The question isn’t simply:
How much press coverage do we have?
A better question is:
What does our press coverage establish about us?
If you sell cybersecurity software but most of your coverage is about funding rounds, your founder or your company culture, you can have an impressive press book without establishing a strong association with the commercial problems you actually solve.
You need category-relevant authority.
The sources talking about your company should connect your brand with the topics, problems, products and categories you want to own.
Muck Rack’s research also shows that the type of question affects the sources AI systems cite. Industry-trend questions generated more than twice as many journalism citations as how-to queries in its analysis.
That means modern PR strategy needs to consider more than the publication name.
Think about the relationship between:
Brand → Category → Problem → Source → Prompt
If you want to appear when somebody asks:
“What are the best enterprise cybersecurity platforms?”
there should be credible material across the web connecting your company with enterprise cybersecurity.
If most of your coverage says you raised $40 million, the web has plenty of evidence about your financing.
It may have much less evidence about why somebody should buy your product.
3. Your competitors have stronger category association
This is one of the biggest differences between traditional brand awareness and AI visibility.
A company can be well known without being strongly associated with a particular category or problem.
Your customers might know you.
Investors might know you.
Journalists might know you.
But when an AI system has to answer:
“What are the leading companies in this category?”
It needs information that helps connect companies with that category.
Think about category association as a web of corroborating evidence.
Your website says you belong in the category.
Industry publications say you belong in the category.
Review sites classify you in the category.
Journalists discuss you alongside other companies in the category.
Comparison articles include you.
Experts mention you.
Customers discuss you in the context of the problem.
Relevant websites repeatedly connect your company, product and category.
Over time, that creates a much clearer digital footprint.
There is some interesting research supporting this.
Ahrefs analyzed 75,000 brands across ChatGPT, Google AI Mode and AI Overviews and found branded web mentions had a strong correlation with AI visibility, ranging from roughly 0.66 to 0.71 across the platforms it studied. Mentions on YouTube showed an even stronger correlation.
An earlier Ahrefs study specifically examining AI Overview visibility also found that brands in the top quartile for web mentions averaged substantially more AI Overview mentions than brands with fewer mentions.
Correlation is not causation, and CMOs should be suspicious of anyone presenting it that way.
But the direction of the evidence is useful.
Being meaningfully discussed across the relevant web appears to be associated with greater AI visibility.
4. The content AI needs doesn’t exist
Sometimes the problem really is your website.
Growth-stage companies often build websites around conversion.
You get:
Product.
Features.
Customers.
Pricing.
Book a demo.
That architecture can work perfectly well for somebody who already understands the category and knows the company.
It can leave enormous gaps for somebody researching a problem.
Suppose buyers are asking:
- What are the best platforms for managing X?
- How does X compare with Y?
- What should a company look for when buying X?
- What are the alternatives to Company Y?
- What does X cost?
- Which X platform is best for a 500-person company?
- What are the advantages and disadvantages of X?
- How do European and US providers differ?
Does your website actually answer those questions?
More importantly, does it answer them with anything genuinely useful?
Google’s 2026 guide to optimizing websites for generative AI search is unusually clear about this.
Google recommends creating unique, valuable, non-commodity content and says its generative AI features remain rooted in its core Search ranking and quality systems. It also warns against manufacturing pages for every possible prompt variation simply to manipulate AI visibility.
The objective isn’t to manufacture 500 pages for AI.
It is to understand what your customers want to know and become one of the best sources available for answering those questions.
We’ve built our own approach around the same principle. The Proper Propaganda GEO framework moves through six stages: Audit, Strategy, Foundation, Content, Authority and Measurement.
Content is one part of that system.
It isn’t the whole system.
5. AI systems don’t have a clear picture of your company
Here’s another problem we see.
Your company describes itself one way on its homepage.
An industry directory describes it another way.
Your LinkedIn profile is outdated.
Old press releases use positioning you abandoned three years ago.
Journalists describe your product differently.
Your product pages use internal terminology that nobody else in the industry uses.
One source puts you in one category.
Another puts you somewhere else.
Individually, these discrepancies may seem fairly harmless.
Collectively, they create a messy representation of what the company actually is.
For GEO, we want the important facts surrounding a company to be easy to discover and reasonably consistent:
Who are you?
What do you sell?
Who is it for?
What category do you belong to?
What problems do you solve?
What products do you offer?
How are those products different?
Technical foundations still matter here.
Google recommends maintaining a clear technical structure, crawlability and conventional SEO best practices for generative AI search. It specifically says there is no special schema or AI-specific markup required for appearing in AI Overviews or AI Mode.
Structured data can still be useful as part of normal SEO.
But there isn’t a magic GEO schema that fixes weak authority or confused positioning.
6. The sources influencing AI aren’t talking about you
This is where GEO starts looking a lot like very good PR.
You need to understand which sources actually appear around the questions that matter in your category.
There is no universal list.
For one market it might include TechCrunch, Forbes, YouTube and Reddit.
For another it could be a collection of specialist trade publications, review sites, industry blogs and expert sources that most consumers have never heard of.
Muck Rack’s analysis found AI systems cited journalism from more than 20,000 outlets. It also found that the publications commonly considered “top tier” by PR teams weren’t necessarily the publications most frequently cited by AI systems in individual industries.
That’s important.
The traditional PR question is:
Which publications reach our customers?
GEO adds another:
Which sources influence the AI answers our customers see?
Sometimes those lists overlap.
Sometimes they don’t.
Finding that difference should influence media strategy.
It is also why we look at citation analysis as part of a broader GEO program rather than treating every media placement as equally valuable.
7. Your competitors have a stronger citation footprint
Your ChatGPT problem may ultimately be a competitive authority problem.
A competitor might have:
- more relevant editorial coverage
- more appearances in comparison articles
- more product reviews
- more category-specific mentions
- more expert commentary
- stronger YouTube visibility
- stronger participation in relevant communities
- better explanatory content
- clearer product information
- stronger associations with the problems customers are researching
One or two mentions aren’t particularly interesting.
The footprint is.
If your competitor repeatedly appears across the sources surrounding a category while you appear sporadically, AI systems simply have more information available about them.
This is why GEO shouldn’t be measured by asking ChatGPT five questions on a Friday afternoon.
You need to look at the pattern.
How to diagnose why ChatGPT recommends your competitors
Before launching a GEO campaign, establish the problem.
At Proper Propaganda, we start by understanding how a brands customers are using AI and then we beging building a prompt universe and benchmarking the brand against its competitors.
A prompt universe is a structured collection of the questions customers might ask AI throughout the buying journey.
Our GEO Buyer Journey framework maps those questions across six stages:
Education → Discovery → Evaluation → Validation → Purchase → Post-Purchase
That matters because somebody asking:
“How does this technology work?”
has a very different objective from somebody asking:
“What are the best companies offering it?”
And somebody asking:
“Company A vs. Company B?”
has moved further down the buying journey again.
The prompts, information sources and metrics that matter change with them.
The AI visibility diagnostic
If competitors are consistently beating you in AI recommendations, examine these eight areas:
| Diagnostic | What you’re trying to learn |
| Share of Answer | How frequently does your company appear across relevant AI answers compared with competitors? |
| Prompt Coverage | Which high-value buyer questions include your company, and which don’t? |
| Position in Answer | When you appear, how prominently are you surfaced? |
| Recommendation vs. Mention | Does AI actually recommend your company or simply mention it? |
| Citation Analysis | Which sources are being used around important category prompts? |
| Competitor Citation Gap | Which influential sources discuss competitors but not you? |
| Category Association | Which categories, problems and attributes are consistently associated with your brand? |
| Owned Citation Rate | How frequently is your own content used as supporting evidence? |
You can see the full measurement methodology in our GEO framework and measurement guide.
The important part is running these diagnostics against a meaningful set of prompts.
For a B2B technology company, that could include questions such as:
“What companies solve X?”
“What is the best software for X?”
“Company A vs. Company B.”
“What are the best alternatives to Company A?”
“Is Company A reliable?”
“Which X platform is best for enterprise companies?”
“What should I look for when choosing X software?”
Then benchmark your company and competitors against the same prompt universe over time.
Patterns start appearing.
A simpler five-question diagnostic for CMOs
You don’t need a full GEO platform to begin investigating the problem.
Start with five questions.
1. Authority
Are credible independent sources talking about us?
2. Relevance
Are those sources discussing us in connection with the categories and problems we want to own?
3. Coverage
Does useful content exist for the questions our customers actually ask?
4. Clarity
Can search engines and AI systems clearly understand what our company does, what we sell and which market we belong to?
5. Competitive footprint
Are our competitors appearing across more of the sources influencing AI answers?
If you struggle with several of these questions, your recommendation gap becomes easier to explain.
Why PR and GEO are converging
For years, SEO, PR and content were usually managed as separate disciplines.
AI search is making those divisions considerably less useful.
Technical SEO helps machines discover and interpret your website.
Content gives them useful information to retrieve.
PR creates independent evidence elsewhere on the web.
GEO connects those disciplines to the questions customers are actually asking and measures whether the work translates into AI visibility.
Google’s own guidance reinforces the continued importance of strong technical foundations and useful, original content.
Muck Rack’s citation research points to the importance of earned sources across ChatGPT, Claude and Gemini.
Ahrefs’ research points toward the relationship between wider brand mentions and AI visibility.
Taken together, an interesting picture emerges.
Your website matters.
What the rest of the web says about you matters too.
That’s one reason we combine PR and GEO rather than treating them as unrelated marketing functions.
What should a CMO do if competitors dominate ChatGPT recommendations?
Don’t start by publishing 100 AI-optimized blog posts.
Start with the evidence.
1. Build your prompt universe.
Identify the questions prospects ask from early education through purchase.
2. Establish your baseline.
Measure where your company currently appears across those questions.
3. Identify the competitors beating you.
Look for patterns rather than individual answers.
4. Map the citation environment.
Determine which publications, websites, communities and other sources repeatedly appear around important prompts.
5. Compare your authority footprint.
Where are competitors being discussed that you aren’t?
6. Identify your content gaps.
Which important customer questions don’t have a genuinely strong answer on your site?
7. Fix your brand clarity.
Make your positioning, category, products and differentiators clear and consistent.
8. Build authority where the gaps exist.
That could mean editorial coverage, product reviews, comparison inclusion, expert commentary, original research, useful content or participation in relevant communities.
9. Measure again.
GEO is an optimization loop.
The objective is to see whether changes in your website, content and authority footprint correspond with changes in how AI systems understand and surface the brand.
That’s the logic behind our broader GEO methodology at Proper Propaganda.
The real question isn’t whether ChatGPT knows your company
Most growth-stage technology companies are asking the wrong question.
They ask:
“Does ChatGPT know who we are?”
I’d ask something more useful:
“Does the web contain enough credible, relevant and consistent evidence for AI systems to recommend us when a buyer asks about our category?”
That’s a much higher bar.
Your website needs to explain what you do.
Your content needs to answer the questions customers actually ask.
Your technical infrastructure needs to make that information accessible.
And credible sources outside your website need to substantiate the things you say about yourself.
If competitors keep appearing in AI recommendations and you don’t, somewhere in that system there is probably an evidence gap.
Find the gap first.
Then fix it.