How Does AI Find Your Brand? Latest Findings for PR and Marketing

Your next visibility opportunity might be a better service page, a useful how-to guide, or an interview with a trade publication.

Muck Rack’s latest research gives comms, PR, and marketing teams a reason to consider all three.

Muck Rack’s October 2026 findings show that about 84% of AI citations come from earned media under its broad definition. But the details matter: journalism’s share has declined, owned content serves a distinct purpose, and different AI platforms draw on substantially different sources. The research and methodology are available through Generative Pulse, a Muck Rack company.

My takeaway: don’t choose between your website and PR. Build a strategy that connects what you publish, what others say about you, and the questions your audience asks.

In my earlier article, AEO in 2026: Why Earned Media and Your Website Both Decide What AI Says About You, I explored why both channels matter. This follow-up focuses on the latest findings and what I recommend doing with them.

Which Sources Do AI Answer Engines Cite?

Muck Rack’s October 2026 research found that about 84% of citations came from broadly defined earned media, including journalism, research, government sources, social content, and third-party corporate content. Within owned-media citations, product and service pages accounted for 58%. The practical takeaway: connect useful website content with relevant third-party coverage rather than treating them as separate visibility strategies.

Key Takeaways

• Earned media accounts for about 84% of citations in Muck Rack’s research, but that category extends well beyond journalism.
• Journalism represents about 22% of citations, down from approximately 27% in May 2026. That decline largely shifted citations into other earned-media categories.
• Your website still has a clear role. Among owned-media citations, product and service pages account for 58%, followed by practical guides at 24%.
• Audience intent shapes the content opportunity. How-to questions favor practical guides, best-of questions favor buyer’s guides and reviews, and industry-trend questions favor expert analysis.
• One answer isn’t an audit. Repeating the same question can produce different sources, so measurement needs consistent prompts, repeated runs, and platform-specific reporting.

Research Profile: How Muck Rack Studied AI Citations

Before applying the findings, it helps to understand what the research measured.

Muck Rack’s October 2026 research analyzed more than 40 million AI responses across two complementary datasets. This was an observational analysis of AI answers and their cited sources—not a survey of consumers, marketers, or journalists.

Methods and coverage vary by dataset and chart; not every finding represents all 40 million responses.

For example, the September provider comparison used 1,020 prompts per provider. A separate longitudinal comparison repeated the same 102 questions from April to September. Keeping those analyses distinct helps avoid treating every percentage as if it came from one identical test.

Providers also introduced and retired models during the research period. Changes over time can therefore reflect model updates as well as changes in source selection.

My interpretation: use the findings to guide your content, PR, and measurement decisions—not as a guarantee that a particular page or placement will produce an AI citation.

What Does “84% Earned Media” Actually Mean?

It does not mean 84% news coverage.

Muck Rack’s earned-media definition includes:

• Journalism.
• Academic research.
• Government and NGO sources.
• Encyclopedic sites.
• Social platforms and user-generated content.
• Third-party corporate blogs and content.

That last category is especially useful to understand. A company’s article can be classified as third-party corporate content when that company is not the brand targeted in the question. Content from the company being asked about is classified as owned media.

The accurate headline is that broadly defined earned sources account for about 84% of citations. Journalism alone accounts for about 22%.

This is also citation research, not a measurement of everything an AI model knows. It doesn’t establish that 84% of model training comes from earned media or that PR controls 84% of an answer.

For communicators, I see a practical implication: think beyond a traditional media list. Relevant research, specialist publications, useful third-party resources, and credible community discussions belong in the visibility conversation.

Journalism’s Share Declined. Does PR Matter Less?

Not necessarily. The more useful question is what role journalism plays.

In the October research, 48% of journalism citations point to breaking or recent news, while another 30% point to buyer’s guides and reviews.

That makes journalism valuable for timely developments and outside evaluation. It can explain what changed, compare options, and put an organization’s claims into context.

Meanwhile, owned-media citations lean heavily toward product and service pages and practical guidance.

Those are complementary jobs:

• Your website explains your offering and approach.
• Independent coverage adds reporting, evaluation, and perspective.
• Together, they give audiences more than one way to understand your expertise.

My recommendation is to stop treating coverage and website content as separate campaigns. When you earn attention, make sure the pages people visit next accurately explain the work behind it.

Product and Service Pages Account for 58% of Owned-Media Citations

AEO discussions can quickly become conversations about writing more blog posts.

The latest findings suggest taking a closer look at your core business pages first. Product and service pages represent 58% of owned-media citations in the research; practical guides represent another 24%.

For a consultant, association, speaker, or professional services firm, I recommend reviewing whether your pages clearly answer:

1. Who do you help?
2. What problem do you solve?
3. What does the engagement or program include?
4. How does your approach differ?
5. What evidence or examples support your expertise?
6. What should an interested reader do next?

A clever headline can open the door. The rest of the page needs to explain what’s inside.

For my own strategic advisory work through CoryWest Media, that means explicitly connecting AI visibility, PR, leadership positioning, and advisory frameworks like The Expert Gap™ through clear descriptions—not expecting readers to assemble the relationships themselves.

Match Your Content to the Question Intent

One of the most useful findings is that different questions favor different formats.

These figures describe content formats within query categories, not the likelihood that any individual page will receive a citation.

Here’s how I translate those patterns into a content strategy:

• For How-To Questions, Teach: Create useful instructions, examples, checklists, and explanations. Answer the question fully enough that a reader can act.
• For Best-Of Questions, Support Evaluation: Make your offering easy to understand and compare. Provide verifiable information, then pursue relevant opportunities for independent reviews or buyer guidance.
• For Industry Questions, Contribute a Perspective: Offer analysis grounded in your experience, research, or observations. Explain what changed, who it affects, and why it matters.
• For Comparison Questions, Explain the Differences: Address tradeoffs directly. Help readers understand which approach fits which situation, rather than declaring one universal winner.

The strategic shift is simple: start with the audience’s question, then choose the format.

Different AI Platforms Draw on Different Sources

A strong result on one platform doesn’t tell you how you appear everywhere else.

In Muck Rack’s September collection of 1,020 prompts per provider:

• ChatGPT drew 35% of classified citations from government and NGO sources.
• Gemini drew 34% from owned media.
• Claude’s largest source category was third-party corporate content at 24%.
• Perplexity had the highest journalism share among the four providers, at 28%.

Only 212 domains appeared across all four providers in that collection.

These are observed patterns, not permanent platform identities. They do support a diversified approach: useful owned content, relevant earned coverage, and reporting that keeps platforms separate.

Don’t Overlook Trade and Regional Publications

The research showed that Claude, Gemini, and Perplexity cited smaller news outlets than ChatGPT. It specifically highlighted Claude’s use of trade and regional publications.

For PR teams, that’s a reason to evaluate relevance alongside reach.

A specialist publication that understands your industry may be a stronger fit for your story than a national outlet with little connection to your audience. Build your media list around that fit, then examine which outlets appear across repeated AI visibility checks.

One AI Answer Is Not a Visibility Audit

You ask a question. Your business appears. Great.

You ask it again. The sources change.

Muck Rack measured cited-domain overlap when the same prompts were run twice on the same day:

These figures measure source overlap between runs—not answer accuracy or the probability of a brand appearing.

The lesson is to treat a single answer as an observation, not a verdict.

Stable prompt sets, repeated runs, provider-specific reporting, tracking ranges/rolling trends, and documenting model changes are essential for accurate measurement. It is also critical to track brand mentions separately from citations.

That last distinction matters: AI can mention your organization without linking to your site. A linked page and a favorable description are different outcomes.

My 5D WIRED PR Approach to These Findings

I teach visibility through my 5D WIRED PR Framework: Digital, Direct, Dynamic, Data, and Distinctive. Here is how to apply those dimensions to Generative Engine Optimization:

Digital: Make Your Expertise Easy to Understand

Review your website’s service pages, biographies, resources, and program descriptions. Make sure they reflect your current work and answer useful questions.

Direct: Prioritize Audience Relevance

Choose outreach opportunities because the story fits the publication and its readers—not simply because the outlet has a recognizable name.

Dynamic: Connect Your Content and Coverage

Turn a substantive idea into useful formats: a practical article, an expert interview, a presentation, or a research commentary. Give each format a purpose rather than repeating the same promotional message.

Data: Measure Patterns, Not Screenshots

Track whether you appear, how you’re described, and which sources are cited. Compare repeated results by platform.

Distinctive: Make Your Point of View Recognizable

Use consistent descriptions of your expertise, supported by specific examples and original thinking. The goal isn’t identical wording everywhere; it’s a coherent picture of what you know and why it matters.

A Practical AI Visibility Action Plan

1. Choose the Questions That Matter: Build a manageable set of questions your customers, members, buyers, or event planners actually ask. Include how-to, comparison, recommendation, industry-trend, and risk questions—not only searches for your name.
2. Establish a Repeatable Baseline: Run the same prompts more than once across the platforms relevant to your audience. Record the date and model when available.
3. Track Mentions, Citations, and Accuracy: Use a structured worksheet to log the prompt, platform, brand appearance, expertise description, cited sources, mentioned competitors, and missing context. A favorable mention isn’t the same as an accurate one.
4. Improve Your Most Important Pages: Start with pages explaining your services, expertise, programs, and practical guidance. Use query intent findings to decide where you need a guide, a comparison, a stronger service description, or original analysis.
5. Connect the Gaps to Your PR Plan: Look for questions where useful third-party context is missing. Pursue relevant interviews, research contributions, trade coverage, and independent evaluation.
6. Review Trends and Adjust: Conduct a quarterly strategic review, with more frequent checks during major campaigns or website changes.

Frequently Asked Questions

What Is the Difference Between AEO and GEO?

Answer engine optimization and generative engine optimization are overlapping terms used in work aimed at understanding and improving visibility in AI-generated answers. Muck Rack’s report discusses both in connection with how platforms portray and recommend brands. In this article, AEO refers to that broader visibility work.

Does Muck Rack’s 84% Figure Mean News Coverage?

No. Its earned-media category includes journalism, research, government and NGO sources, encyclopedic sites, social content, and third-party corporate content. Journalism alone accounts for about 22% of citations in the October report.

Does My Website Still Matter for AI Visibility?

Yes. Among owned-media citations, product and service pages account for 58%, followed by practical guides at 24%. Those findings support giving your core pages and instructional content a clear role in your strategy.

Can Trade and Local Media Contribute to AI Visibility?

Yes. The report documents citations to smaller publications and highlights Claude’s use of trade and regional outlets. That supports considering specialist and regional coverage rather than focusing exclusively on large national publications.

How Can I Check Whether AI Describes My Business Accurately?

Use a consistent set of audience questions, repeat them across platforms, and review both the descriptions and cited sources. Track brand mentions separately from citations and look for patterns over time, following the report’s measurement guidance.

Can an AEO Strategy Guarantee Recommendations?

No. This research documents variation across questions, platforms, model versions, and repeated runs. It can inform your decisions, but it doesn’t establish a guaranteed route into an AI answer.

What Does AI Say About Your Business?

Is it accurate? Does it reflect your expertise today? Which sources shape the answer?

My AI Visibility & AEO Audit examines how major AI platforms describe your brand, which sources they cite, and where your website and public messaging leave important questions unanswered.

Through CoryWest Media, I help executives, associations, and marketing teams turn those findings into a practical visibility roadmap.

Work with Barbara Rozgonyi to connect your expertise, PR, and digital presence.

About Barbara Rozgonyi

Barbara Rozgonyi is a keynote speaker, AI and PR strategist, and founder of CoryWest Media. Through her 5D WIRED PR Framework and Brighter Presence™ platform, she helps leaders and organizations connect their expertise, messaging, and visibility. She publishes WIRED PR Works and works with executives, associations, and marketing teams through strategic advisory, AI visibility audits, keynotes, and workshops.

Welcome! I’m glad you’re here! Tell me what’s on your mind, whether that’s an event you’re planning, a brand that needs a fresh set of eyes, or a question you’re not sure who else to ask. Looking forward to learning more about you and your project.

Ask Barbara about an AI Visibility & AEO Audit

Get Connected 

LinkedIn | YouTube | Instagram | Facebook | X | Flickr | Podcast | Newsletter