The AI visibility gap: Why paid, owned channels are no longer enough for B2B tech

Around 50% of internet users hop on genAI tools as part – or all - of their research; Google searches per user in the US dropped ~20% year-on-year (2024–2025), and Gartner predicts a 25% drop in traditional search volume by 2026.

So where does AI get its information from, and how can you ensure you are part of that source?

New research from media management platform Muck Rack, titled Generative Pulse, provides one of the clearest answers to date - and the implications are huge for how companies and brands think about their approach to marketing, PR, content and visibility in an AI-driven environment. 

The report's objective was to quantify and characterize the nature of AI-generated citations across different use cases and vendor models. This included frequency, source types, and the prominence of earned and owned media. Gemini, Perplexity, Claude and ChatGPT were used to execute the queries, between July and December 2025. And yes, the research is US-centric, but so are those models.

AI is an earned media machine

The most striking finding is that around 94% of links cited by AI models come from non-paid sources, with earned/news media the largest contributor, at around 25%. That is followed by third-party/corporate blogs (24%), aggregators/encyclopaedic (15%), first-party/corporate blog (12%), and academic, government, social, each representing 4%-7%.

Why is that? In simple terms, when AI tools generate answers and reference sources, almost everything they rely on is drawn from:

• News articles
• Independent and third-party blogs
• Research papers and academic content
• Government and institutional sources
• Reference platforms such as encyclopaedic sites – Wikipedia, etc

That means earned, and other non-paid channels, are the primary pathway into how AI understands and references brands, and so how it delivers answers and recommendations to you.

Think about that. If your business marketing is relying on paid, your website, and your socials – which many companies do – you are significantly limiting your visibility in AI-generated answers – and the audiences you need to reach. 

That is because for AI tools, trusted news outlets – Reuters, Bloomberg, Straits Times - signal trust to AI models, with authority and verification driving ranking, and traditional media shaping digital knowledge systems. AI anchors itself - at least in part - to credible media outlets and journalism.

Freshness is a ranking factor - but velocity matters even more

AI models also show a strong preference for recent content. The report notes that 50% of citations come from content published within the last 11 months, with the highest citation rates occurring within the first seven days of publication

This creates a very different dynamic from traditional SEO. It is not just about producing evergreen content. It is about sustained publishing cadence, timely commentary and rapid amplification of news and insights. In practical terms, companies that move quickly - especially around news cycles - are more likely to shape AI outputs.

That does present its own challenges. When I asked ChatGPT about my company, it was probably 90% accurate, and drew on some recent thought leadership, which, while not inaccurate, somewhat skewed the information. And so monitoring GenAI on an ongoing basis is as important as monitoring social and news channels.

Authority is not generic - it is contextual

AI models do not just prioritise "high authority" sites in general. They prioritise authority within a specific domain or industry. The research shows that models cite globally recognised outlets, but also rely heavily on industry-specific publications for industry specific queries.

This has two implications. First, there is no universal "top media list" that works for every brand, and second, specialist and vertical media remain highly valuable. AI models draw from distinct ecosystems depending on the industry being queried. The research shows that about 50% of AI citations for a brand can come from as few as 20 outlets. However, those outlets are different for every brand, are shaped by industry and topic, and are built over time through consistent visibility. This reinforces the need for precision in media strategy, not just scale.

For B2B companies in particular, this reinforces the importance of sector relevance over simple audience reach. Below are examples of how AI sources information differently for different business sectors.

Technology: Technology queries draw from a narrower set of dominant outlets, with fewer unique sources cited compared to other industries. This creates a more concentrated and competitive media environment.

Finance / Insurance: A mix of financial media, advisory content and consumer platforms dominate, balancing authority with accessibility.

Education: AI leans toward institutional, academic and specialist sources, reflecting the structured, research-led nature of the sector.

Energy: Coverage is shaped by regulatory, policy and industry analysis, with strong representation from government and specialist publications.

The bigger shift: AI is reshaping the value of communications

Taken together, these findings point to a broader shift. AI is not just another channel. It is becoming a synthesis layer - one that aggregates information, filters for authority, and prioritises relevance and recency. In that system, visibility is no longer driven by ad spend, content volume and owned media. Instead, it is driven by credibility, consistency and contextual authority, which are dominated by non-paid information sources.

What this means for companies in Southeast Asia

For companies across Singapore and Southeast Asia - particularly in B2B technology - how you market and communicate is changing in important ways.

First, PR/media is fundamental communications infrastructure. Second, content must extend beyond your owned channels. Third, industry-specific media – business, finance, technology, etc - matters more than reach alone. Fourth, speed and consistency shape AI visibility, and finally authority must be built, not assumed.

When it comes to adapting to GenAI and company communications, firms need to shift from asking "How do we appear in AI?" to "Are we producing the kind of information AI trusts?" You may not be, but you need to. AI does not create authority, but it reflects it very powerfully.

*The author, Patrick Keenan, does not use, represent or endorse Muck Rack.