AI Search Visibility Report for August 2026: LinkedIn enters the top three

Contents

• August confirms a more divided AI citation landscape

• Top individual sources: Citation volume and trends

• LinkedIn reaches the top three

• YouTube and Reddit retain the top two positions

• Social performance splits sharply by platform

• Earned media falls below 30%

• Reference sources continue to lose ground

• The model divide gets wider

• AI citations reach further into specialist media

• What PR and marketing teams should do next

• Conclusion: Build a portfolio of citation paths

• FAQ

August confirms a more divided AI citation landscape

Generative AI platforms are changing how people discover companies, compare products, and form an opinion before they ever visit a brand’s website. For PR, communications, and marketing teams, that makes the sources behind an AI answer almost as important as the answer itself.

Using Meltwater GenAI Lens, we analyzed thousands of prompts and approximately 7.3 million citations across eight major AI platforms: Anthropic Claude, ChatGPT, Copilot, Google AI Mode, Google AI Overviews, Google Gemini, Perplexity, and xAI Grok. This report compares August with July and places the results within six months of tracking from March to August 2026.

The prompt set covers a balanced mix of high-intent consumer categories, including AI, retail banking, automotive, fast food, streaming services, apparel, and fashion. Prompts were distributed equally across categories and contained no brand mentions. All month-over-month figures come from the platform’s August versus July calculations.

Source preferences can vary with prompt wording, category, location, timing, and the specific version or retrieval system behind each AI platform, which is why repeated measurement is more useful than a one-off snapshot.

August builds on the pattern we saw in July’s AI search visibility report, although the picture has become more uneven. LinkedIn rose again, driven mainly by Perplexity. YouTube and Reddit remained the two most-cited individual domains, even as the overall social category declined. At the same time, Earned/News, Wikipedia, NIH, and Statista all lost ground.

The result is a more fragmented citation landscape. Different models are drawing from very different information ecosystems, and even sources within the same category are moving in opposite directions.

Top individual sources: Citation volume and trends

YouTube, Reddit, and LinkedIn formed the top three in August. LinkedIn was the only one of those sources to increase its citation volume, while Yahoo was the only other top-five domain to grow.

The ranking is important, though the movement across several metrics gives us a fuller view. LinkedIn increased in citation volume, citation rate, and visibility score together. Reddit’s volume held nearly flat while its citation rate and visibility score improved. YouTube declined in volume after July’s rebound, yet its citation rate and visibility score stayed level.

LinkedIn reaches the top three

LinkedIn has produced the clearest multi-month rise in the current dataset. It moved from eighth in June to fifth in July and third in August, reaching 100.2K citations after another 24.8% month-over-month increase.

Perplexity drove most of that growth. It generated 69,302 LinkedIn citations in August, roughly 69% of LinkedIn’s total within this sample. That is the largest single model-source pairing recorded across the six months, narrowly ahead of the 68,073 Gemini citations and 68,011 Google AI Mode citations generated by YouTube in August.

LinkedIn is growing across the overall dataset, although much of its current scale depends on one model. Google AI Mode cited LinkedIn 13,855 times, Google AI Overviews 8,232 times, Copilot 3,941 times, Grok 3,540 times, and Claude 699 times. ChatGPT and Gemini recorded zero LinkedIn citations among the tracked data.

For B2B brands, LinkedIn now deserves deliberate attention within a GEO plan. Company research, executive expertise, practical explainers, and clear points of view can create useful public source material, especially when a post states its main point early and supports it with specific evidence. Our report on how LinkedIn content wins in AI search explores that opportunity in more detail.

The model breakdown also sets a clear guardrail: LinkedIn activity should support a broader source strategy. Its influence varies widely from one AI platform to another.

YouTube and Reddit retain the top two positions

YouTube remained the leading individual source with 212.5K citations, despite a 7.34% decline from July’s record 229.3K. Its citation rate and visibility score were unchanged, which suggests that YouTube’s reach across the tracked prompt set remains broad.

This was a softer month after July’s strong rebound, rather than a second consecutive decline. YouTube also remained well above its May total of 188.9K citations. September’s data will tell us whether August was normal variation or the start of a longer downward run.

The model data continues to show why YouTube holds such a large lead. Gemini cited it 68,073 times, Google AI Mode 68,011 times, Google AI Overviews 38,217 times, and Perplexity 30,774 times. For brands, clear titles, detailed descriptions, accurate transcripts, chaptering, and focused educational content can give AI systems more useful information to extract.

Reddit stayed in second place with 127.0K citations, down just 0.95%. Its citation rate improved 16.7% and visibility score increased 5.56%, so the underlying picture was stronger than the small volume decline suggests. Gemini generated 48,972 Reddit citations, Perplexity 35,007, Google AI Mode 20,684, and Google AI Overviews 16,146.

That broad support across several models makes Reddit more resilient than any one headline about a single AI platform might suggest.

Social performance splits sharply by platform

At category level, Social declined from 7.8% in July to 7.2% in August. That puts it level with April and still above March, May, and June. The aggregate number hides four very different platform stories: LinkedIn grew, Reddit remained almost flat, Facebook reversed its July increase, and Instagram declined for a second month after its June surge.

Facebook fell 29.6% to 62.4K citations, giving back the gain it recorded in July. Instagram dropped 43.1% to 30.3K and showed declines in citation rate and visibility score as well. The June surge across those platforms has not developed into a sustained upward run.

This volatility makes channel-level measurement essential. A strategy based only on the overall Social share could miss LinkedIn’s rise, Reddit’s resilience, and the sharp reversals on Facebook and Instagram.

The “Other” category needs equally careful interpretation. Its share grew to 61.3%, yet its rounded count remained at 4.5 million, the same as July. The category covers a wide range of corporate sites, research databases, market intelligence platforms, financial reference sites, and other web sources. Its rising share shows that sources beyond the named categories dominate the mix, while the stable rounded count gives us no evidence that every type of owned content increased in August. Brands should track citations to their own domains directly instead of treating “Other” as a proxy for owned-content performance.

Earned media falls below 30%

Earned/News accounted for 29.4% of citation domain count in August, down from 33.1% in July. This is the third consecutive monthly decline, following 37.6% in May and 35.0% in June, and the first reading below 30% in the six-month dataset.

That downward run deserves attention. Earned media still contributed approximately 2.1 million citation domain counts, more than any category except the broad “Other” group. High-quality reporting continues to provide third-party validation, category context, expert analysis, and source diversity across AI answers.

For communications teams, the practical response is to connect earned coverage with other citable sources. A strong media story can support an executive LinkedIn post, a YouTube explainer, a research page, a customer-facing FAQ, or a clearly sourced article on the company website. Each format gives AI platforms another route to the same credible information.

Press Release held at 0.2% for a fourth month. Press releases continue to serve disclosure, distribution, and media relations needs, while the direct AI citation opportunity remains small in this dataset.

Reference sources continue to lose ground

Three sources that rose earlier in the year declined together again in August. NIH fell 19.1% to 65.6K citations, its second monthly decline. Wikipedia dropped 14.8% to 56.8K, and Statista fell 15.9% to 42.5K, giving both sources a third consecutive decline.

They still sit inside the top 10, so their authority has not disappeared. ChatGPT and Claude continued to cite NIH heavily, Copilot favored Statista among the ten leading domains, and Grok cited Wikipedia more than any other domain in the table.

The data supports maintenance over abandonment. Keep public factual information accurate. Publish clearly sourced data where it serves the audience. Correct outdated information and make supporting evidence easy to find. Then assess new investment against the models, topics, and source pathways that influence your own brand.

The model divide gets wider

Aggregate rankings can make the citation environment look more consistent than it is. The leading domain among August’s overall top 10 changes across the eight AI platforms.

Citations were spread more evenly across institutional, social, and data sources

Perplexity’s profile changed most clearly. LinkedIn led at 69,302, while Reddit reached 35,007, Yahoo 31,433, and YouTube 30,774. Facebook, which led Perplexity in July, fell to 8,994. Anyone measuring Perplexity through a blended AI visibility score would miss that reordering.

Gemini continued to lean heavily toward video and community content within these leading domains. YouTube and Reddit together generated 117,045 citations, while Wikipedia remained a meaningful third source at 22,077. Google AI Mode showed greater breadth, combining very high YouTube volume with Facebook, Reddit, LinkedIn, Yahoo, and NIH.

Claude, ChatGPT, and Copilot retained more data-led or institutional source profiles. Grok remained the most evenly spread across the ten domains. These patterns describe citation behavior within the current sample; they do not measure answer quality or tell consumers which platform is best for a particular type of question.

The industry discussion around Reddit in August shows why those distinctions count. Axios reported on a separate dataset in which Reddit’s share of ChatGPT Search citations fell from an average 3.83% through August 7 to 0.52% by August 14. Our tracked sample recorded zero ChatGPT citations to Reddit from May through August, even as Reddit held second place across all eight platforms.

Different prompt sets, collection windows, and methodologies can produce different baselines. Both findings reinforce the same operating principle: examine each model on its own and keep monitoring, because a source can remain influential across the wider AI ecosystem while disappearing from one platform’s citations.

AI citations reach further into specialist media

August also brought considerable turnover among the most-cited journalists and content creators. Six new names entered the top 10, spanning sports technology, healthcare, travel, retail, golf, and biotechnology.

Mark Haranas held first place for a fourth consecutive month, showing the continued citation value of enterprise technology reporting. Dorothy Neufeld remained second, although her count declined from 5.7K in June to 2.3K in August.

The arrival of MyGolfSpy, Xtalks, Athletech News, and Becker’s Hospital Review shows how far AI citation pathways can extend into specialist publishing. Query relevance and subject depth can give a focused outlet considerable reach, even when it has less general-market awareness than a large national publisher.

One month of turnover cannot establish a permanent trend, though it gives media relations teams a useful prompt. Journalist mapping for GEO should look beyond the largest publications and identify the specialists whose reporting directly answers the questions audiences ask AI.

What PR and marketing teams should do next

1. Build LinkedIn as a deliberate citation surface

LinkedIn’s third consecutive rise and its 69,302-citation pairing with Perplexity make it a current priority, especially for B2B organizations. Give executive and company content a clear job. Answer specific category questions, state the useful point early, support claims with evidence, and link to deeper source material when appropriate.

2. Keep YouTube and Reddit in the source portfolio

YouTube still leads by more than 85K citations, and Reddit remains second with support from Gemini, Perplexity, and Google’s AI experiences. Maintain accurate transcripts and detailed descriptions on YouTube, then monitor the real questions and language people use in relevant Reddit communities. Genuine participation and useful information will serve the audience regardless of a short-term change in one model.

3. Extend earned stories across other formats

Earned/News has declined for three months, yet it still represents close to three in ten citation domain counts. Use strong media coverage as the credible center of a wider content system. Carry its evidence into owned explainers, social posts, video, executive commentary, and sales materials, with clear attribution back to the original reporting.

4. Measure every model separately

Perplexity’s LinkedIn concentration, Gemini’s YouTube and Reddit volume, and ChatGPT’s institutional profile would disappear inside one aggregate score. Track brand visibility, source influence, sentiment, citation volume, and competitor performance by model, then use the blended view for executive context.

5. Review trends monthly and annotate major platform changes

Citation behavior can change quickly. A monthly review helps teams distinguish a temporary spike from a sustained movement, while an annotated timeline connects performance changes with model launches, retrieval updates, campaign activity, media coverage, and new content.

GenAI Lens brings AI source and citation tracking together with wider media and social intelligence, helping PR, communications, and marketing teams understand how brand narratives travel across the platforms shaping discovery.

Conclusion: Build a portfolio of citation paths

August makes the fragmentation of AI visibility harder to ignore. LinkedIn rose into the top three, mostly through Perplexity. YouTube and Reddit remained the largest individual sources, while Facebook and Instagram fell sharply. Earned/News dropped below 30%, and several reference sources declined again.

No single channel can carry a GEO strategy across every model. Brands need a connected source portfolio that combines credible earned coverage, expert voices, useful owned content, clearly structured video, relevant professional content, community participation, and accurate public reference information.

The mix will keep changing. Teams that measure each AI platform, understand the sources behind its answers, and respond with coordinated content and communications decisions will be better prepared to protect visibility and build authority as generative discovery develops.

FAQ

What was the biggest AI citation development in August 2026?

LinkedIn delivered the strongest sustained rise, growing 24.8% to 100.2K citations and moving into third place. Perplexity generated 69,302 of those citations, the largest model-source pairing recorded in the six-month dataset.

Is earned media becoming less important for AI visibility?

Earned/News share declined for a third consecutive month and reached 29.4%, its lowest level in the dataset. It still generated approximately 2.1 million citation domain counts and remained the second-largest category, so it continues to provide significant reach and third-party authority. The data supports a more connected source strategy around earned coverage.

Why did the overall Social share fall when LinkedIn grew?

The platforms moved in different directions. LinkedIn rose 24.8% and Reddit held almost flat, while Facebook fell 29.6% and Instagram fell 43.1%. Those declines pulled the overall Social share down from 7.8% to 7.2%.

Should brands use the same GEO strategy for every AI platform?

Each platform showed a different source profile in August. Perplexity leaned heavily toward LinkedIn, Gemini toward YouTube and Reddit, Claude and ChatGPT toward institutional sources, and Google AI Mode drew from a broader mix. Brands should track and plan by model, then coordinate those activities within one wider communications strategy.

What should teams watch in September’s data?

Watch whether YouTube declines again, whether Earned/News finds a lower floor, and whether Yahoo follows its 17.1% August rise with another month of growth. Reviews also fell after several stable months, while LinkedIn’s heavy dependence on Perplexity makes its model mix worth tracking closely.