Google just pushed a dedicated Discover Core Update, and if Discover traffic is part of your distribution diet, you should treat this like a platform algorithm shift, not a tiny SEO tweak. The announcement is the cleanest place to start: Search Engine Land’s report on the update.
This update matters to creators for one reason: Discover is increasingly AI-shaped. Not just ranked by AI, but presented by AI, with Google rolling out AI-generated summaries in Discover in 2025. That means the old playbook (spicy headline + trend-jacking + vibes) gets less reliable, while signals like local relevance, topical consistency, and clean structure get more valuable.
Google’s stated direction for this update is pretty straightforward: surface more locally relevant content, reduce sensational or clickbait-style presentation, and elevate in-depth, original, timely work from sources that show topic-specific expertise. In practice, that’s going to reshuffle who wins in the feed, especially for publishers and creator-led sites that live on mobile discovery spikes.
Translation: Discover is getting pickier, and it’s doing it at machine speed.
What changed in Discover
Google hasn’t dropped a single one weird trick ranking factor (sorry). Instead, this is a bundle of distribution adjustments that collectively push Discover toward higher-trust, higher-context content.
Local relevance turns up
The biggest visible shift: more emphasis on content from sites based in a user’s country. Coverage that’s created in-market, clearly relevant to that market, and written with local context is positioned to get a boost.
This isn’t just for local news. If you cover entertainment, tech, lifestyle, or sports, you can still win here, as long as the piece reads like it understands the audience’s world. Generic posts that could have been published from anywhere may still perform, but they may be less likely to get the same Discover lift outside their home market.
Why this is happening: Discover is trying to be useful in a feed environment, not just interesting. And usefulness is often location-shaped: pricing, availability, events, language, cultural references, and what’s actually trending where the user lives.
Clickbait gets de-powered
Google is also pushing harder against sensational packaging, especially headlines that overpromise, obscure the point, or lean too hard on curiosity gaps.
That doesn’t mean you need to write boring headlines. It means the headline needs to accurately preview the value. If your title sounds like it’s hiding the ball, the system is increasingly likely to treat it like it’s hiding the ball.
The practical effect: creators who built Discover reach on mystery headline + fast rewrite tactics should expect volatility.
Expertise gets more specific
The third major shift is how Google evaluates expertise: more topic-by-topic, less overall site vibe.
A generalist publisher can still win, but the model is increasingly asking: Does this source consistently deliver on this subject? A site that reliably publishes strong work in a narrow lane may outrank a bigger brand dabbling in that lane once a month.
This is one of the most creator-friendly parts of the update. You don’t need to be a mega-site. You need to be predictably good at something.
AI summaries change the game
Discover isn’t just selecting your content, it may be summarizing it directly in the feed, which changes how attention flows.
TechCrunch reported that Google began rolling out AI-generated summaries in Discover in July 2025, showing an AI-written overview with multiple publishers’ logos and an AI disclaimer that summaries can make mistakes (TechCrunch). That feature introduced a new reality: users can get the gist without clicking, and publishers compete not only for ranking, but for being included in the summary mix.
This February update fits that direction. When AI is doing more of the front-of-house work, structure and clarity stop being nice-to-have and start being distribution infrastructure.
Here’s what tends to map well to AI summaries:
- Clear subheads that signal what’s happening
- Early context that answers why should I care
- Specificity over vague hype
- Strong attribution and transparent sourcing
And here’s what gets risky fast:
- Long, meandering intros that delay the point
- Slang-heavy writing without clear nouns (fun for humans, confusing for machines)
- Jargon piles without definitions
- Everything posts that never land a conclusion
If your post can’t be summarized cleanly, it’s harder to feature cleanly.
Who benefits (and who doesn’t)
This update doesn’t pick winners by team human vs team AI. It picks winners by whether your output looks like it was made for readers, not for the feed.
Likely to gain
- Local-first publishers with real on-the-ground context
- Niche specialists who publish consistently within a topic cluster
- Creator sites with a recognizable beat and repeatable formats
- Teams who can update posts quickly with new context (without rewriting history)
Likely to lose
- Sites leaning on headline tricks to manufacture curiosity
- High-volume content mills publishing broad, shallow coverage
- Publishers trying to rank in every category without depth
- Posts that are hard to parse (for readers and summarizers)
The new Discover checklist
Not a guide, not a course, not a 47-step Notion template. Just the realities creators should operationalize right now.
Headline honesty wins
Write titles that are still clickable, but fully legible. If your headline implies a payoff, deliver it quickly. If it’s a reaction, say what the reaction is. If it’s an analysis, name the subject.
A good mental test: Would someone feel tricked after reading the first two paragraphs? If yes, that’s now an actual distribution risk.
Local context isn’t optional
If your content has any geographic angle, make it explicit:
- country or region context
- availability differences
- local stats, sources, or examples
- the so what for that audience
This is also where smart automation helps: not to generate spammy city pages, but to adapt framing responsibly.
Topical clusters beat random spikes
If you’re building a creator media business, Discover increasingly rewards the boring thing that works: showing up repeatedly in the same lane.
That doesn’t mean repeating yourself. It means building a body of work that signals: we cover this, we know this, we update this.
Google’s broader guidance on creating helpful, reliable, people-first content is still the north star (Google Search Central), and it lines up cleanly with where Discover is heading.
If you want a COEY-side read on how Google is packaging information with AI right now, see: Gemini 3 Flash Agentic Vision That Double-Checks Images.
At-a-glance impact table
| Update theme | What it favors | What gets weaker |
|---|---|---|
| Local relevance | In-market context, clear geo signals | Generic works everywhere framing |
| Anti-clickbait | Specific, accurate headlines | Curiosity gaps, overpromises |
| Topic expertise | Consistent coverage in a niche | Broad, shallow category hopping |
| AI summaries | Clean structure, fast clarity | Messy intros, vague writing |
What to watch next
Expect two things at once: short-term turbulence and long-term normalization.
In the near term, creators who relied on Discover surges should watch for swings as the rollout settles. Not because Google is punishing anyone, but because Discover is a finite feed and the selection math changed. If new sources get surfaced (more local, more specialized, more structured), somebody else has to move down.
In the longer term, this update reinforces a clear platform direction: Discover is becoming a tighter loop between ranking, summarization, and user satisfaction. And that loop doesn’t love fluff.
The creators best positioned for this shift aren’t the loudest. They’re the ones who can publish with speed and keep their work grounded: real expertise, real context, and headlines that don’t act like a scammy movie trailer.
Discover still rewards momentum. It’s just getting better at spotting what deserves it.
