Two mechanisms now compete for the same real estate in AI answers: Google's Preferred Sources, a user-curated allowlist inside AI Overviews and AI Mode, and digital PR link building, the earned coverage that already accounts for most AI citations. Marketers keep folding both into one bucket labeled "getting cited by AI." They are not the same problem. One is a setting a searcher flips in their own account. The other is a placement you earn from a journalist. Confuse the two and you'll spend a link building budget chasing a lever you don't actually control.
What Google's Preferred Sources actually does
Preferred Sources launched inside classic Search and Google News, letting a signed-in user hand-pick outlets whose stories get bumped up whenever they search a topic. On May 27, 2026, Search Engine Land reported that Google had extended the feature into AI Mode and AI Overviews, the two generative surfaces where more research now starts before a user ever sees a blue link. The mechanic didn't change. A user opens a topic, adds the outlets they already trust, and Google leans on that list when it assembles an AI answer.
The scale is real. Google product manager Duncan Osborn told Search Engine Land that users have curated more than 345,000 unique sources through the feature, and that clicks to a preferred source land at roughly twice the historical click rate of a non-preferred result. Google paired the expansion with two related changes: a perspectives carousel that surfaces forum and social commentary alongside traditional sources, and a wider rollout of highly cited labels that flag sources AI Mode and AI Overviews already lean on heavily. Search Engine Land's reporting on the rollout is worth reading in full if you manage AI visibility for a brand.
None of that is small. A 2x click multiplier is the kind of number a growth team would kill for anywhere else in the funnel. But read the mechanism carefully before you build a strategy around it. Preferred Sources is opted into by an individual user, inside their own Google account, one outlet at a time. You cannot buy a spot on it, you cannot pitch your way onto it at scale, and you get no visibility into who has added you or why. It rewards brand recognition a user already has. It does not build that recognition in the first place.
It's worth naming the closest thing to a lever here, because it's easy to conflate with paid influence and it isn't the same thing. The perspectives carousel and the highly cited labels Google shipped alongside this expansion both reward sources that already show up often and get engaged with, which loops back into brand recognition rather than around it. There is no application, no submission form, and no account manager on Google's side you can talk to about getting added. The only path in is being the kind of outlet or brand a searcher already half-remembers, which is a downstream effect of coverage you've already earned, not a new channel in its own right.
Digital PR link building still drives most AI citations
Earned media didn't wait for Google's product roadmap to become the dominant path into AI answers. Muck Rack's data, which we broke down in rebuilding a PR target list around AI-cited journalists, found that 82% of AI citations trace back to earned media: journalist coverage, expert quotes, and data-led placements, the kind of reporting no platform toggle controls. Paid and owned content combined make up a small fraction of what's left.
That share holds because of how the underlying mechanism works, not because of a quirk in one dataset. A generative engine builds an answer by retrieving pages it already trusts and citing whichever best support a claim. A well-reported article, a data-backed quote, a named byline at an outlet with a citation history: all of that reads as a trust signal to a model the same way it reads as credibility to a human reader. It's the same logic behind the six link types AI engines actually trust: the links and mentions that move citation rates are earned ones, not the kind you generate by publishing more owned pages.
Earned media also isn't a Google-only asset. A placement that gets picked up, cited, and re-cited compounds across ChatGPT, Perplexity, Claude, and Google's own AI surfaces at once, because every one of those engines crawls and re-indexes the same published article. Preferred Sources, by contrast, is a setting inside one company's product. It has no equivalent inside ChatGPT or Perplexity today, and there's no guarantee it survives Google's next redesign in a form that still resembles the list 345,000 users just spent time curating.
Digital PR link building vs. Google's Preferred Sources
Put the two levers side by side and the differences stop looking cosmetic.
| GOOGLE'S PREFERRED SOURCES | DIGITAL PR / EARNED MEDIA | |
|---|---|---|
| Who controls it | The individual user, inside their Google account | You, through pitching, reporting, and relationships |
| How you earn it | Being recognizable enough that a user adds you manually | Original data, expert access, and working journalist relationships |
| Cost to build | Zero direct cost, but no scalable path in | Real budget and time; compounds if reinvested |
| Durability | One Google product change from resetting or disappearing | An owned asset; the article and its citations persist |
| Cross-engine reach | Google surfaces only: AI Mode, AI Overviews, News, Search | Crawled and re-cited across ChatGPT, Perplexity, Claude, and Google alike |
| Current scale | 345,000+ curated sources, 2x click rate (Google, May 2026) | 82% of all AI citations (Muck Rack) |
Two rows matter more than the rest. Durability and cross-engine reach are the whole argument. A Preferred Sources list lives entirely inside a Google product surface that Google can redesign, rename, or fold into something else at any point, the way search features have churned for two decades. An earned placement, once published, doesn't depend on any one platform's roadmap. It gets crawled, indexed, and cited by every engine that trusts the outlet it ran in, which is why an 82% figure like Muck Rack's can hold up across engines, not just inside Google.
The verdict: where to put a limited budget
The reasoning comes down to what each dollar buys. A dollar spent through our link building and digital PR service buys a placement that keeps earning citations across engines for as long as the piece stays live and gets referenced elsewhere. A dollar spent trying to influence Preferred Sources buys nothing, because there's no purchase mechanism to spend it on. The only lever you have is being the kind of source a practitioner in your category already wants to add: consistent bylines, cited data, a reputation that precedes the click. That's a downstream effect of good PR, not a separate initiative worth its own line item.
This is also why we don't build clients a standalone "Preferred Sources strategy." There's no scalable input on the other end of it. What you can do is make the brand-recognition case easier: publish under real, named experts, keep a consistent presence in the outlets your buyers already read, and let generative engine optimization services handle the technical side of making sure engines can find, parse, and re-cite that coverage once it exists. The category-level effect of enough earned coverage is that more of your target buyers already know your name well enough to add you themselves.
We saw the compounding half of this play out directly on a recent engagement: our work with TechTrust rebuilt a thin PR target list around the journalists and outlets actually showing up in AI citations for their category, and the resulting coverage kept generating citations months after the original news cycle closed, the kind of return a one-time Preferred Sources add could never produce on its own. It's also worth watching the platform-level tension here. The same growing reliance on curated and AI-summarized answers is part of why publishers are reconsidering their relationship with Google over AI Overviews, which is one more reason not to build a strategy around a single company's evolving product surface.
There's a reasonable objection worth answering directly: if Preferred Sources drives a real 2x click lift, isn't ignoring it leaving money on the table? Not quite. You're not choosing whether to accept those clicks, you're choosing where to spend effort trying to influence an outcome. Earned coverage moves both numbers at once, since a well-placed story both earns citations directly and makes your brand more likely to be the recognizable name a searcher adds to their own Preferred Sources list later. Chasing Preferred Sources on its own moves neither, because there's nothing to chase.
Do this next
Skip the temptation to chase Preferred Sources with outreach, and don't wait for Google to explain how the list actually gets built. Instead, run the earned-media math this week: pull the last two quarters of your coverage, check how much of it reads as genuinely earned versus paid or syndicated, and compare that against the 82% benchmark. If your earned share is well under that, the fix isn't a new platform tactic. It's redirecting next quarter's link building budget toward original data, expert sourcing, and journalist relationships that compound, the same rebuild we walk through for PR target lists, and letting any Preferred Sources benefit show up as a side effect of doing that well.
See where you are cited today
A free snapshot audit of your rankings and AI citations before we ever talk.
Josh leads work at the intersection of SEO and generative engines at Something Inc., helping B2B brands get ranked and cited across every major AI engine.