How to Increase Brand Mentions in AI: RankCaster AI Data from 8,058 Observations

A RankCaster AI white paper

Executive Summary

How can a brand move from being merely discoverable by AI systems to being mentioned more frequently in AI-generated answers?

We examined this question across four projects that published content through the RankCaster AI Content Manager. Over six weeks, we tracked how owned content, external mentions, and citation frequency influenced the appearance of brands in AI-generated responses.

In total, we collected 8,058 observations. Each observation represents the result of checking one target prompt for one brand during one week.

The study found that:

• without citations, brands were mentioned in only 0.5% of responses;
• owned content alone increased the mention rate to 35%;
• external signals alone produced a mention rate of approximately 40%;
• combining owned content with external signals increased the rate to 58%;
• when a brand received 6-20 citations per week, the average mention rate reached 50%;
• with 21 or more citations per week, the mention rate reached 61%.

The central finding is clear:

Increasing AI visibility requires more than publishing content on a brand’s own website. A brand must also appear consistently in relevant third-party sources, including social platforms, directories, partner websites, and other independent publications.

---

What We Studied

Most AI visibility tools answer the question:

Where is AI already mentioning the brand?

This is an important diagnostic metric. It shows which pages are cited, which competitors appear in AI responses, and which topics already generate visibility for the brand.

However, monitoring alone does not explain:

• which actions increase the frequency of brand mentions;
• how much content a brand needs to publish;
• what role external sources play;
• why the same brand appears for one prompt but not another;
• what a decline in mention rate means;
• when newly published content begins influencing AI responses.

We therefore examined two related events:

1. An AI system cites a source associated with the brand.
2. The AI system mentions the brand in the text of its response.

This allowed us to move from a purely descriptive question, “Where does the brand appear?”, to a more actionable one:

“Which content and visibility signals increase the likelihood that AI will mention the brand?”

---

Research Methodology

From August to September 2026, we tracked four brands that published content through the RankCaster AI Content Manager:

• a marketing agency in Israel;
• a mortgage broker in Israel;
• a contractor-matching platform in the United States;
• an education center in Poland.

The platform checked target prompts daily across multiple AI systems. For each response, we recorded:

• whether the brand’s URL was used as a source;
• whether the brand itself was mentioned in the response;
• which external sources appeared alongside the brand;
• how many citations the brand received each week;
• how the brand’s mention rate changed over time.

Over the six-week period, we collected 8,058 observations. Each observation represents the result of checking one target prompt for one brand during one week.

This structure allowed us to compare both different brands and changes within the same project, before and after content publication, during periods of increased citation activity, and when competitors became more visible within the same topic.

---

Two Types of Visibility Signals

We divided citation sources into two categories.

Owned signals

Owned signals occur when an AI system uses the brand’s own domain as a source.

Examples include:

• blog articles;
• service pages;
• case studies;
• research reports;
• expert content;
• FAQs;
• analytical articles.

Owned content helps AI systems understand:

• what the brand does;
• which services it provides;
• which topics it covers;
• what type of expertise it has;
• which audience and market it serves.

External signals

External signals occur when an AI system cites a third-party website that mentions the brand.

Examples include:

• social media posts;
• business directories;
• industry publications;
• partner websites;
• rankings and aggregators;
• cross-brand content;
• pages published by other companies.

External sources create independent context around the brand. They indicate that the brand exists not only on its own website, but also within the broader ecosystem of its market.

---

Owned Content and External Mentions

The results were distributed as follows:

Signal combination	Observations	Mention rate
No signals	3,277	0.5%
Owned content only	1,925	35%
External signals only	2,905	40%
Owned content and external signals	851	58%

In this sample, external signals alone produced a slightly higher mention rate than owned content alone: 40% versus 35%.

The strongest result came from combining both signal types: 58% mention rate.

This does not mean that external mentions can replace a brand’s own website. Without high-quality pages on the brand’s domain, AI systems have less information about the brand’s services, positioning, and expertise.

The reverse is also true: publishing articles without building external context may not generate maximum visibility.

A more complete strategy requires both:

• owned pages that explain the brand and its subject matter;
• external sources that reinforce the brand’s presence in that subject area.

---

Which External Sources Performed Best?

Cross-brand mentions

A cross-brand signal occurred when one brand mentioned another brand from the same client portfolio.

For two Israeli projects, reciprocal mentions generated:

• 664 citations for one brand;
• 698 citations for the other.

This was the largest individual source of external signals in the study.

For agencies, this effect can become an additional portfolio asset. When several clients operate in related categories, relevant mentions between them can expand each brand’s presence across AI sources.

However, cross-brand mentions should be natural and contextually justified. The brands should be connected by a shared topic, service, market, or user need. Mechanical link exchanges without a meaningful relationship do not create genuine topical authority.

Social media

Facebook, Instagram, and LinkedIn generated approximately 100-120 citations per brand during the observation period.

Their contribution was smaller than that of owned content, but social platforms created a consistent background signal.

Social content is more useful when it:

• uses a consistent brand name;
• describes specific services;
• relates to priority topics;
• links to relevant pages on the website;
• appears regularly in a professional context.

Directories and aggregators

The data included:

• Yelp;
• Clutch;
• G2;
• Alignable;
• SourceForge.

Each individual source generated between 1 and 9 citations. This is a small contribution on its own, but the combined effect helps create a broader picture of the brand.

For AI systems, the relevant factor is not only the number of mentions on one platform. It is the overall pattern: the brand appears across multiple independent sources in connection with the same category or topic.

---

Citation Frequency and Brand Mentions

After combining owned and external signals, we grouped observations according to the number of brand citations recorded during each week.

Brand citations per week	Observations	Mention rate
0	3,277	0.5%
1-5	1,250	21%
6-20	2,214	50%
21+	1,011	61%

The largest difference appeared between two groups:

• 0 citations, 0.5% mention rate;
• 6-20 citations, 50% mention rate.

The difference was 49.5 percentage points.

The same pattern appeared independently across all four projects: as the number of relevant brand citations increased, the likelihood of a brand mention also increased.

This table should not be interpreted as a fixed formula. Six citations do not automatically produce a 50% mention rate for every brand. Results are also affected by:

• market competition;
• content quality;
• prompt relevance;
• language and region;
• indexing speed;
• competitor activity;
• AI system differences;
• page freshness.

The figures are best understood as planning and diagnostic ranges.

---

What Happened in Each Project?

Marketing agency, Israel

• Mention rate: 40.5%.
• Average citations from the brand’s own domain: 8.4 per week.
• At 37 citations per week for one prompt, the brand appeared in 100% of responses.
• After citations dropped to 9-12 per week, the mention rate for the same prompt fell to 17-20%.
• External signal: 664 cross-brand citations from a portfolio partner.

This example shows that the intensity of a brand’s presence matters even when the topic and source remain relevant.

Mortgage broker, Israel

• Mention rate: 41.7%.
• Number of tracked prompts: 13.
• Average citations: 7.6 per week.
• External signal: 698 cross-brand citations.

Visibility remained stable, with the domain continuing to appear among the sources used in AI responses.

Contractor-matching platform, United States

Content publication began on August 9.

Period	Mention rate	Citations per week
Before publication	6.3%	0.5
After publication	14.4%	2.5

After publication began:

• the mention rate increased by 8.1 percentage points;
• the relative increase was 130%.

However, 2.5 citations per week remained within the lowest 1-5 citation range. The brand achieved a clear improvement, but had not yet reached the level associated with a 40-50% or higher mention rate in this sample.

Moving into the 6-20 citation range required approximately three to eight times more content or external signals, depending on content quality and topic competitiveness.

Education center, Poland

The first article was published on August 26.

Within three weeks:

• citations increased from 13 to 65 per week;
• for the prompt “SMM courses in Europe,” 12 citations corresponded to a 50% mention rate;
• at 18 citations, the mention rate reached 67%;
• for the prompt “SEO courses,” the mention rate fell to 0% after citations stopped.

This example shows how strongly prompt-level relevance can influence AI visibility. A page does not need to be broadly visible for every topic; it needs to be highly relevant to the specific questions the brand is trying to win.

---

Why AI Visibility Can Decline

High visibility does not remain stable automatically.

For one project, the prompt “Russian-speaking event advertising agency” continued to generate 10-17 citations per week, but the mention rate declined:

• from 100% at the beginning of the observation period;
• to 43% after six weeks.

The number of citations did not fall. This indicates that the brand had not disappeared from the source set.

A likely explanation is that competitors increased their presence around the same topic. AI systems continued to cite the brand’s pages, but mentioned other companies more often in the final response.

What to do when mention rate declines

When citations remain stable but mention rate falls, the team should:

1. identify which competitors have appeared in the responses;
2. compare their content, wording, and source profiles;
3. update outdated pages;
4. add new facts, examples, and proof points;
5. increase relevant external mentions;
6. rerun the same prompt after the content is updated.

A decline in mention rate is not necessarily a strategy failure. It is a signal that the competitive environment has changed and the content strategy needs to adapt.

---

What Can 3-4 Articles per Month Achieve?

Based on the projects in the study, we developed the following indicative model:

Scenario	Indicative mention rate
3-4 articles per month on the brand’s own domain	~35%
External signals only	~40%
Owned articles and external signals combined	~58%

These are directional figures, not guaranteed outcomes. Results depend on the market, topic, quality of the content, and how closely each publication matches the monitored prompts.

Potential timeline

Weeks 1-2

The first articles are published and begin the indexing process. A visible effect may not appear immediately. A lag of 7-14 days is possible.

Weeks 3-4

The first pages may begin appearing among AI sources. Individual prompts may generate 5-20 citations and a mention rate in the 20-50% range.

Month 2

With approximately 5-8 accumulated articles, citation volume may reach 15-40 per week. The indicative mention rate is 40-55%.

Month 3

With approximately 9-12 accumulated articles, citation volume may reach 25-60 per week. The indicative mention rate is 50-60%.

Results depend especially on two factors.

Market competitiveness

The Polish education market produced 65 citations after the publication of one article. By contrast, the U.S. home repair topic generated approximately 2.5 citations per week despite a substantial volume of content.

Prompt alignment

An article must address the same question represented by the monitored prompt. Including similar keywords is not enough if the page does not answer the underlying user intent.

---

The RankCaster AI Visibility Cycle

RankCaster AI identifies prompts
with low brand visibility
          ↓
Content Manager creates a topic
for a specific information gap
          ↓
The article is published
on the client’s domain
          ↓
The page is indexed by AI systems
and search crawlers
          ↓
AI begins using the page
as a source for relevant answers
          ↓
External signals develop in parallel:
partners, social media, directories,
and cross-brand mentions
          ↓
AI sees the brand across multiple
sources related to the same topic
          ↓
Mention rate is measured again
          ↓
Content is updated or reinforced

---

Preparing Content for SEO and GEO

For pages intended to be understood by both search engines and AI systems, teams should:

• begin with a direct answer to the main question;
• maintain one primary topical focus;
• write headings around real audience questions and tasks;
• use the brand name and key entities consistently;
• include verifiable facts, dates, and figures;
• provide specific examples;
• connect the page to other relevant content on the website;
• update information when the market changes;
• use structured data only when it accurately reflects the visible page content.

Google recommends creating helpful, reliable, people-first content and using descriptive, specific page titles. Structured data can help search engines understand page content, but it does not guarantee enhanced search visibility or citation by AI systems. developers.google

For GEO, the practical priorities are:

• clear definitions;
• direct answers;
• consistent brand descriptions;
• verifiable claims;
• a clear relationship between the brand, its service, and the topic;
• regular content updates;
• visibility across multiple independent sources.

Google also recommends concise article headlines because long headlines may be truncated on some devices. developers.google

---

What This Means for Agencies

AI visibility is not only a question of what a brand publishes on its own website.

A complete strategy should address two areas:

1. which pages are created and updated on the brand’s domain;
2. where the brand is mentioned in connection with priority topics.

Owned content provides the primary explanation of the brand. External sources add independent context and reinforce the association between the brand and a particular category, service, or market.

In our sample:

• owned content alone produced approximately 35%;
• external signals alone produced approximately 40%;
• the combined approach produced 58%.

The portfolio effect

An agency managing several brands in related categories can create an additional portfolio effect:

• one client is mentioned in another client’s relevant content;
• publications from different brands reinforce the broader topical network;
• external signals are distributed across several projects;
• AI systems receive more context connecting each brand to its target category.

These mentions should be natural and useful. Their purpose is not to create a formal link network, but to explain genuine relationships between brands, services, and audience needs.

---

Measuring the Outcome

For each priority prompt, teams should track:

Metric	What it shows
Mention rate	How often AI names the brand
Citation rate	How often AI uses a brand-related source
Owned citations	Contribution from the brand’s own domain
External citations	Contribution from third-party sources
Weekly citations	Accumulated volume of visibility signals
Competitors in responses	Which brands are gaining ground on the same topic
Degradation	Where visibility declines despite continued citations
Indexing lag	How long it takes for publication to influence results

Recommended workflow

1. Select commercially important prompts with low mention rates.
2. Check whether the brand already has a relevant page.
3. Create or update content around the specific information gap.
4. Connect the page to other relevant content on the site.
5. Reinforce it with relevant external mentions.
6. Measure both URL citations and actual brand mentions.
7. Update the strategy when new competitors appear or performance declines.

---

Conclusion

The RankCaster AI study analyzed 8,058 observations across four brands over six weeks.

The results showed:

• no signals: 0.5% mention rate;
• owned content only: approximately 35%;
• external signals only: approximately 40%;
• owned and external signals combined: up to 58%;
• 6-20 citations per week: approximately 50%;
• 21 or more citations per week: approximately 61%.

The central finding is:

AI visibility can be measured and improved through a systematic combination of owned content, external mentions, and strong topical relevance.

Most tools show where a brand is already appearing. RankCaster AI helps connect visibility data to specific actions:

• which topic to target;
• which page to create;
• which content to update;
• where to strengthen external presence;
• how to measure the impact of publication;
• when to respond to declining visibility.

The strategic question is no longer:

“Does AI mention our brand?”

It becomes:

“What should we do to make AI mention our brand more often for commercially important topics?”