Last updated: July 21, 2026 · LeadsNow AI, Melbourne
For twenty years, getting found online meant one thing: build links. Guest posts, directories, outreach emails, anchor-text spreadsheets. That playbook just quietly died — not because Google changed, but because your buyers moved. They now ask ChatGPT, Perplexity and Google’s AI Overviews who to hire, and those engines don’t care about your backlink profile. They care about who is talking about you. The data on this is now overwhelming, and it points in one uncomfortable direction: the citations that decide whether AI recommends your business overwhelmingly come from media you don’t own.
AI answer engines overwhelmingly cite earned media — third-party journalism, industry publications and review sites — not brand-owned blogs. Muck Rack’s Generative Pulse study of 25 million AI-cited links found earned media accounts for 84% of all AI citations, while paid and advertorial content gets just 0.3%. For B2B service businesses, the fastest route into AI answers is digital PR: original data, expert commentary and third-party coverage.
The data: AI engines cite other people’s coverage of you, not your blog
Muck Rack’s Generative Pulse “What Is AI Reading?” study (May 2026 edition) analysed more than 25 million links cited in ChatGPT, Claude and Gemini responses across 17 industries. The headline findings:
- Earned media accounts for 84% of all AI citations. Across all three editions of the study since July 2025, the figure has ranged from 82% to 89% — it is not a blip, it is the structure of the system.
- Journalism alone makes up 27% of cited sources.
- Paid and advertorial content accounts for 0.3%. You cannot buy your way into an AI answer with sponsored placements.
The December 2025 edition, built on more than one million links cited by ChatGPT, Claude, Gemini and Perplexity, put it even more bluntly: about 94% of all citations come from non-paid sources, with earned media alone at 82% in that dataset. In other words, when an AI assistant tells a buyer which lead generation agency, mortgage broker or fit-out firm to consider, roughly nine times out of ten it is quoting something a third party wrote.
That flips the content-marketing default on its head. Most B2B service businesses spend nearly all their “visibility” budget on their own blog and their own socials — the channel category that earns the smallest share of AI citations. The businesses getting recommended are the ones being written about.
Owned vs earned vs paid: how AI engines treat each
| Channel | Share of AI citations | How answer engines treat it | What actually works |
|---|---|---|---|
| Earned media (journalism, industry blogs, reviews, forums) | ~84% (Muck Rack, May 2026) | Default evidence base. Treated as independent verification of who’s credible. | Original data journalists want, expert commentary, inclusion in roundups and comparison articles. |
| Owned content (your blog, resource pages) | Minority of citations, but not zero | Cited when it’s the primary source — unique benchmarks, definitions, structured answers nobody else publishes. | Answer-first pages, proprietary numbers, frequent updates. Also does the converting once the click arrives. |
| Paid placement (advertorials, sponsored posts) | ~0.3% (Muck Rack, May 2026) | Effectively invisible. Engines route around content that reads as advertising. | Almost nothing, for citation purposes. Spend it on PR or original research instead. |
Old link building vs 2026 digital PR
Digital PR isn’t “link building with a nicer name”. The target changed. Link building optimised for an algorithm that counted links; digital PR optimises for language models that quote coverage. The practical differences:
| Old link building (2005–2023) | Digital PR for AI citations (2026) | |
|---|---|---|
| Goal | A followed link with the right anchor text | Your brand named, quoted and contextualised in text an AI will retrieve |
| Currency | Domain authority of the linking site | Whether the publication is in each engine’s actual citation pool |
| Asset | Guest post, infographic, directory entry | Original data, benchmark studies, expert commentary, review-site presence |
| Measurement | Referring domains, rankings | Share of answer — how often AI answers name you for your money queries |
| Failure mode | Penalty for manipulative links | Simply never appearing in the conversation buyers actually have |
One PR hit won’t cover you: every engine reads a different internet
Here’s the complication most agencies gloss over: the engines barely overlap. An Averi analysis found that only 11% of sites get cited by both ChatGPT and Perplexity — and that 40–60% of citations change month to month. Getting quoted in one publication that ChatGPT loves does not mean Perplexity or Gemini will ever surface it.
5WPR’s AI Platform Citation Source Index 2026, which synthesised 680 million citations from six major studies (August 2024–April 2026), maps how skewed the terrain is:
- Reddit is the #1 source across every major AI engine, cited at roughly 40% frequency across LLMs.
- Wikipedia accounts for 26% to 48% of ChatGPT’s top-10 citation share.
- The top 15 domains capture 68% of all consolidated AI citation share — concentration beyond anything classic PageRank produced.
- Journalism’s share jumps to 49% on time-sensitive queries, up from 27% overall — news coverage matters most exactly when buyers are actively comparing options.
- It’s volatile: ChatGPT’s Reddit citation share fell from roughly 60% to 10% in six weeks in late 2025 after a single Google parameter change.
The takeaway for a B2B service business: spread your earned footprint across source types — trade press, industry blogs, review platforms and community threads — rather than betting everything on one channel. We’ve broken down the two most distinctive engines separately: how B2B brands get cited in Perplexity and how Reddit feeds AI search citations.
How a B2B service business earns AI citations (without a PR retainer the size of a salary)
1. Publish original data nobody else has
Journalists and industry bloggers cite numbers, and AI engines cite the journalists. If you run any volume of client work, you’re sitting on benchmarkable data — response times, conversion rates, cost patterns, seasonal trends. Package it once a year with a clear methodology and you become the primary source that earned coverage quotes.
2. Be the quoted expert, not the guest poster
Contributed commentary — a sharp quote in a trade-press piece, a sourced answer to a journalist request — puts your name inside earned media, which is where 84% of the citations live. One caution from Muck Rack’s December data: there was only about 2% overlap between the journalists PR teams pitch most and the journalists AI actually cites. Chase the publications that show up in AI answers for your queries, not the ones with the most prestige.
3. Get into the roundups and comparison listicles
“Best X in Australia” and “Top agencies for Y” articles are citation magnets because they map exactly onto how buyers phrase questions to AI. Being one entry in a credible third-party list frequently beats an entire category of self-published content.
4. Don’t write off press releases — write them differently
Muck Rack’s December 2025 edition found press release citations grew 5x between July and December 2025. The releases that get cited look like reference material, not hype: on average they carry 2x more statistics and 2.5x more bullet points than uncited ones. Announce data, not adjectives.
5. Keep everything fresh
AI engines punish stale sources. AirOps’ stale-content research found 83% of citations for commercial queries come from content updated within the past year, 60% from content updated in the last six months, and pages left over a year without updates are more than twice as likely to lose their citations. Muck Rack saw the same pattern from the other side: half of all AI citations pointed to articles published within the previous 11 months. Earned coverage decays too — digital PR is a drumbeat, not a launch.
Book a call if you’d rather talk through what this means for your pipeline than build a PR function from scratch — we’ll show you how we approach it for our own category.
Where owned content still earns its keep (and how we practise this ourselves)
None of this means “stop publishing”. It means owned content has a different job. Your pages get cited when they’re the primary source — original benchmarks, precise definitions, direct structured answers — and they do the converting once a buyer clicks through from an AI answer. Earned media wins the mention; owned content wins the meeting.
We’ll be honest about our own approach, since this article is part of it. LeadsNow publishes original benchmark and answer-first content in our niche, keeps it updated, and measures the result with a single metric: share of answer — how often AI assistants name us when buyers ask about pay-per-result lead generation. That measurement discipline matters more than any single tactic, because with 40–60% of citations churning monthly, you cannot manage what you check twice a year.
The commercial backing behind the content is the part no PR can fake: 50,769+ AI-booked sales appointments since 2017 and 1M+ leads generated, documented in 25 filmed client case studies, with a 4.6★ rating across 43 Google reviews and clients ranging from Foundr to Colliers to Sam Tajvidi’s 121 Brokers. Third parties can only write about results that exist — which is why the honest first step of digital PR is doing citable work.
Frequently asked questions
What counts as earned media for AI citations?
Any coverage you didn’t pay for and don’t control: journalism, trade and industry publications, independent blogs, review platforms, community discussions and third-party roundups. In Muck Rack’s Generative Pulse research this category accounts for 84% of AI citations, with journalism alone at 27%.
Does traditional link building still help AI visibility?
Not in its old form. Citation share is extremely concentrated — per 5W’s AI Platform Citation Source Index 2026, built on 680 million citations, the top 15 domains capture 68% of consolidated AI citation share. A directory link from a site no engine cites does nothing; a mention inside a publication that’s already in the citation pool does the work links used to do.
Should we stop publishing on our own blog?
No — change its job. Owned pages still earn citations when they’re the primary source for unique data or the clearest direct answer to a query, and they’re what converts visitors after an AI answer sends them your way. But if your entire visibility strategy is your own blog, you’re competing for the smallest slice of the citation pie.
How do we measure whether digital PR is winning AI citations?
Track share of answer: run your real money queries through ChatGPT, Perplexity, Gemini and Google AI Overviews on a schedule and record how often you’re named and which sources each engine cites. Because only about 11% of domains are cited by both ChatGPT and Perplexity, you have to check each engine separately — one platform’s results tell you almost nothing about another’s.
How often does earned coverage need refreshing?
Continuously. AirOps found 83% of citations on commercial queries come from content updated within the past year, and pages that go unrefreshed for over a year are more than twice as likely to lose citations. Aim for a steady cadence of new data, new coverage and updated pages rather than one annual splash.
Book a call and we’ll walk you through how earned citations plus answer-first content translate into booked, qualified appointments — the metric we actually get paid on.