AI Search Weekly: Publisher Visibility Report

The battle for visibility in AI search is creating new winners and losers every week. As AI platforms like ChatGPT and Perplexity increasingly mediate how millions of users discover information, understanding citation patterns has become essential for publishers navigating this transformative shift. This week's data from ALLMO.ai reveals dramatic platform differences, extreme concentration dynamics, and surprising category-specific patterns that challenge conventional assumptions about AI search visibility.

ALLMO.ai tracks these patterns through systematic testing across diverse prompts, news topics, and autocomplete suggestions, providing publishers with actionable intelligence about their AI search performance. See the full data at allmo.ai/trends. This week's findings highlight just how rapidly the AI search landscape is evolving, and how much strategy matters for maintaining visibility.

Top 3 Highlights

1. The Reuters Effect: Extreme Concentration on ChatGPT-5

Reuters captures an extraordinary 64.1% citation rate on ChatGPT-5, more than double its already-dominant 31.0% performance on ChatGPT-4. This level of concentration suggests winner-take-most dynamics that may fundamentally challenge mid-tier publishers competing for AI visibility. When a single source captures nearly two-thirds of citations, the implications for traffic diversity and publisher sustainability become urgent questions.

2. Platform Divergence: YouTube Dominates Perplexity While Absent From ChatGPT

YouTube achieves 51.2% citation share on Perplexity but has zero presence in ChatGPT's top citations, illustrating fundamentally different content strategies across platforms. Perplexity's multimedia approach creates opportunities for video-first publishers and broadcasters, while ChatGPT's text focus rewards traditional wire services. Publishers can no longer assume a single content strategy will work across all AI platforms.

3. Category Specialists Find Their Niche Despite Overall Concentration

While overall visibility concentrates heavily around Reuters and AP News, category-specific queries reveal opportunities for specialist publishers. TechCrunch achieves 20.9% citation share in startup queries despite just 3.8% overall, and ESPN captures 25.0% in sports despite minimal general news visibility. This demonstrates that topical authority still matters—but publishers must build it strategically within defined verticals.

Overall Platform Leaders

ChatGPT-5

• Reuters dominates with 64.1% citation rate, nearly doubling its ChatGPT-4 performance and achieving unprecedented concentration for a single source

• AP News captures 29.1% citation share, up from 20.6% on ChatGPT-4, cementing the duopoly of wire services on OpenAI's newest model

• The Guardian achieves 16.2% citation rate after minimal presence in ChatGPT-4's top 10, suggesting ChatGPT-5 may value editorial perspective alongside wire service reporting

• Financial Times enters the top tier with 14.5%, indicating premium business journalism gains traction as the model evolves

• Wikipedia drops dramatically to 19.7% from 40.5% on ChatGPT-4, suggesting the newer model prioritizes primary news sources over reference material

Key Takeaway: ChatGPT-5 demonstrates increasing concentration around traditional wire services and premium news brands, with Reuters and AP News together capturing over 90% citation share. The dramatic reduction in Wikipedia citations suggests OpenAI's newest model has been trained or prompted to prioritize primary news sources, creating a narrow competitive landscape where established news organizations dominate and mid-tier publishers face significant visibility challenges.

ChatGPT-4

• Wikipedia leads with 40.5% citation rate, reflecting ChatGPT-4's reliance on encyclopedic reference material for context and factual grounding

• Reuters captures 31.0%, establishing wire service credibility even in the older model architecture

• AP News achieves 20.6%, making wire services collectively dominant even when Wikipedia holds top position

• Democracy Now appears at 8.7%, indicating ChatGPT-4 includes independent and progressive news sources in its citation mix

• Le Monde reaches 6.3%, showing some international and non-English source diversity that becomes less visible in ChatGPT-5

Key Takeaway: ChatGPT-4 shows more diversity in source types, balancing reference material, wire services, and alternative news outlets. The heavy Wikipedia reliance suggests this model architecture treats AI search partially as a knowledge synthesis task, while ChatGPT-5's shift away from Wikipedia indicates a strategic evolution toward primary source journalism.

Perplexity

• YouTube dominates with 51.2% citation rate, reflecting Perplexity's fundamentally different multimedia approach to answering queries

• ABC News achieves 17.1% and CBS News captures 15.5%, showing broadcast networks perform significantly better on Perplexity than on ChatGPT platforms

• Wikipedia maintains 18.6%, suggesting Perplexity balances reference material with other sources rather than abandoning it entirely

• Council on Foreign Relations reaches 9.3%, indicating Perplexity citations favor institutional think tanks and analysis sources

• Deloitte captures 7.8%, showing consulting and professional services firms gain visibility through Perplexity's business-oriented content strategy

Key Takeaway: Perplexity's citation patterns reveal a platform optimized for multimedia and institutional sources rather than pure wire service journalism. The dominance of YouTube alongside strong broadcaster and think tank performance suggests Perplexity positions itself as a research assistant that synthesizes video, institutional analysis, and traditional news—creating opportunities for publishers with diverse content formats.

Category Analysis: Economics & Finance

• Reuters leads with 46.2% citation share, demonstrating wire service dominance in financial news extends across AI platforms

• IMF achieves 25.6% citation rate, showing institutional and governmental economic sources compete effectively with commercial news media

• Wikipedia maintains 25.6%, tied with IMF, suggesting reference material remains valuable for economic concepts and definitions

• AP News captures 20.5%, reinforcing wire service duopoly even in specialized category queries

• Deloitte reaches 12.8%, indicating professional services firms successfully position their research and analysis for AI citation

• Specialized financial publishers like Trading Economics, Financial Content, and T. Rowe Price each achieve 10.3%, demonstrating niche financial data providers gain visibility in targeted queries

Key Takeaway: Economics and finance queries show extreme concentration around wire services and institutional sources, with Reuters and IMF together capturing over 70% of top citations. The presence of professional services firms and specialized data providers indicates that authoritative financial analysis can compete with traditional journalism, but the barrier to entry appears high. Commercial media outlets without wire service scale or institutional credibility face significant challenges in this category.

Category Analysis: Start-ups & Technology

• TechCrunch dominates with 20.9% citation share, nearly 6x its 3.4% overall performance, proving category specialization delivers outsize returns in vertical queries

• Crunchbase News achieves 11.9%, showing database-backed journalism earns AI platform trust through structured, verifiable company information

• StartupBlink captures 10.4% despite minimal overall visibility, demonstrating data aggregators and ecosystem mappers provide value AI platforms recognize

• Reuters maintains 19.4% even in specialist category, showing wire services remain competitive through comprehensive technology coverage

• Wikipedia reaches 17.9%, higher than its ChatGPT-5 overall performance, suggesting reference material stays relevant for technology company backgrounds

• PR Newswire appears at 9.0%, indicating AI platforms cite press releases directly when covering startup funding announcements

Key Takeaway: Technology and startup queries reward specialization and structured data more than other categories, with TechCrunch and Crunchbase capturing over 30% combined citation share through category expertise. The presence of press release services and ecosystem databases suggests AI platforms value verifiable funding data and company information, creating opportunities for publishers who combine editorial judgment with structured business intelligence.

Category Analysis: Sports

• ESPN and Reuters tie at 25.0% and 27.3% respectively, showing both sports specialists and general wire services compete effectively in this category

• Wikipedia achieves 25.0%, likely due to comprehensive athlete biographies, statistics databases, and historical sports records

• YouTube captures 15.9%, reflecting sports highlights, analysis, and commentary video content gaining AI platform citations

• Fox Sports reaches 11.4%, demonstrating broadcast sports networks maintain visibility through comprehensive game coverage

• UEFA, Sky Sports, and Sports Illustrated each achieve approximately 9%, showing official league sources and traditional sports magazines remain relevant

• Specialized publications like Cycling Weekly and FourFourTwo appear at 6.8%, indicating niche sports media can capture citations in sport-specific queries

Key Takeaway: Sports queries show more balanced competition between wire services, specialist outlets, and reference material than other categories, with no single source dominating above 30%. The strong Wikipedia performance suggests AI platforms rely heavily on statistical and biographical data, while YouTube's presence indicates video content plays a significant role in sports information retrieval. This category offers the most opportunities for specialist publishers to compete with general news organizations.

Category Analysis: General News & Politics

• Wikipedia leads with 41.0%, showing reference material dominates when queries span broad topics without clear specialist angles

• Reuters captures 40.0%, nearly tied with Wikipedia, demonstrating wire service authority in general news and political coverage

• AP News achieves 30.0%, making wire services collectively dominant with 70% combined citation share

• YouTube reaches 27.0%, indicating video news and political commentary gain significant traction in general interest queries

• Financial Times appears at 4.0% despite minimal sports or technology presence, showing business journalism maintains relevance in political economy coverage

Key Takeaway: General news and politics queries revert to extreme concentration around wire services and reference material, with Reuters, Wikipedia, and AP News capturing over 110% combined citation share (accounting for overlap). This category offers the least opportunity for specialist or mid-tier publishers, as AI platforms default to the most authoritative general sources when queries lack specific topical focus. Publishers without wire service scale must identify and own specific political or news niches to gain visibility.

Regional Analysis: German-Language Queries

Platform Patterns

• Government sources dominate with Bundesregierung.de at 21.9% overall and Bundestag.de at 18.8%, showing institutional sources outperform commercial media in German-language queries

• Public broadcasters Deutschlandfunk (18.8%) and ZDF (10.9%) achieve strong citation rates, reflecting German public media's authoritative position

• Reuters reaches 12.5% overall but captures 30.0% on ChatGPT-5, showing English-language wire services compete even in German queries

• YouTube achieves 14.1% overall and 40.9% on Perplexity, maintaining its multimedia platform dominance across languages

• German Wikipedia reaches 15.6% overall with 27.3% on ChatGPT-4, showing reference material remains important but less dominant than in English queries

Commercial Media Performance

• Traditional German newspapers like Süddeutsche Zeitung, Focus, and N-TV each achieve 6.3% citation rates, showing commercial media struggles to match government and public broadcaster visibility

• Welt reaches 6.3% overall but captures 20.0% on ChatGPT-5, suggesting premium commercial outlets can compete on newer models

• Spiegel appears at 6.3% with strong Perplexity performance (18.2%), indicating investigative journalism gains traction on specific platforms

• Commercial media collectively captures less citation share than government sources alone, highlighting the structural challenge facing private news organizations in German AI search

Key Takeaway: German-language queries reveal fundamentally different citation dynamics than English, with government and public broadcasting institutions capturing over 40% of top citations while commercial newspapers struggle to reach 10% individually. This pattern suggests non-English markets may see institutional sources dominate AI search more than commercial media, potentially due to training data composition, language-specific trust signals, or the strong public media tradition in German-speaking countries. Commercial publishers in non-English markets face the dual challenge of competing with both English-language sources and domestic institutional media.

Strategic Implications

Multi-Platform Requirements

• Publishers must develop distinct strategies for ChatGPT (wire service competition) versus Perplexity (multimedia and institutional sources) rather than assuming universal approaches

• Video content creation becomes essential for Perplexity visibility, with YouTube citations reaching 51.2% while remaining completely absent from ChatGPT top 10

• Broadcast networks and multimedia publishers should prioritize Perplexity optimization, where ABC News and CBS News achieve 17.1% and 15.5% respectively compared to minimal ChatGPT-5 presence

• Wire services and text-focused publishers maintain advantage on ChatGPT platforms, where Reuters and AP News capture over 90% combined citation share

• Cross-platform visibility requires investment in multiple content formats and distribution strategies, as single-format publishers risk complete invisibility on platforms that don't match their strengths

Key Takeaway: The era of platform-agnostic content strategy has ended for AI search optimization. Publishers must consciously choose which platforms to prioritize based on their content strengths and audience goals, as the gap between ChatGPT's text-focused wire service model and Perplexity's multimedia institutional approach continues widening. Success requires either multi-format capability or strategic focus on platforms aligned with existing strengths.

Specialization Strategies

• Category specialists achieve 5-6x citation rates in their verticals compared to overall performance, with TechCrunch reaching 20.9% in startup queries versus 3.8% overall

• Publishers should identify and dominate 2-3 specific topical areas rather than pursuing general news visibility, where Reuters and AP News concentration makes competition nearly impossible

• Structured data and verifiable information improve citation rates, as evidenced by Crunchbase News (11.9%) and financial data providers (10%+ in economics queries)

• Press release integration provides direct citation opportunities, with PR Newswire reaching 9.0% in technology coverage by offering source material AI platforms cite

• Sports and technology categories offer the most specialist opportunity, while general news and politics remain dominated by wire services

Key Takeaway: Mid-tier publishers must pursue aggressive specialization strategies to compete in AI search, focusing resources on specific categories where topical authority can overcome wire service scale advantages. The data suggests that being the definitive source for startup funding, specific sports, or financial data delivers better AI visibility than attempting broad news coverage. Publishers should consider narrowing focus to 2-3 categories where they can achieve top-3 specialist positioning.

Concentration Challenges

• Winner-take-most dynamics accelerate on newer models, with ChatGPT-5 showing Reuters at 64.1% versus ChatGPT-4's more distributed 31.0%

• The top 3 sources on ChatGPT-5 (Reuters, AP News, Wikipedia) capture over 110% combined citation share, leaving minimal room for additional publishers

• Mid-tier general news publishers risk becoming completely invisible as AI platforms optimize toward fewer, more authoritative sources

• Non-English markets show even higher concentration toward institutional sources, with government sites and public broadcasters dominating German queries

• Publishers outside the top 10 in any category should assume they receive near-zero AI-driven traffic and plan business models accordingly

Key Takeaway: AI search concentration presents an existential challenge for mid-tier publishers who cannot achieve top-3 positioning in specific categories. The data suggests platforms are optimizing toward fewer, more authoritative sources over time, making early positioning critical. Publishers should conduct honest assessments of whether they can realistically achieve top-tier AI visibility or whether alternative strategies (direct audience relationships, social platforms, email) should receive greater investment.

Conclusion

This week's data underscores the rapid evolution and increasing stratification of AI search visibility. The dramatic differences between ChatGPT-5's wire service concentration and Perplexity's multimedia approach demonstrate that AI search is not a unified channel but rather a fragmented landscape requiring distinct strategies. For publishers, success demands either category specialization strong enough to overcome general news concentration, or multi-format capabilities that enable platform-specific optimization. As these patterns continue evolving weekly, ongoing monitoring through tools like ALLMO.ai becomes essential for publishers making strategic decisions about where to invest limited resources in the AI search landscape.