Generative Engine Optimization (GEO) and the New Shape of PR
Did AI and generative engine optimization (GEO) make press releases cool again?
There was a moment, not long ago, when many startups proudly declared they didn’t "do PR." Press releases were dismissed as expensive and outdated, replaced by founder updates and blog-first content strategies.
To be fair, it worked for a while, and these tactics can still be effective depending on your audience and objective. But the landscape is shifting again, and this time it’s not just about distribution, it’s about visibility at its core.
That’s why questions like “Should this go on the wire?” or “Do we pitch this to a journalist or post it on LinkedIn?” miss the bigger point. These questions focus on where content goes, not how it earns authority, credibility, and presence in the new search environment shaped by AI.
As I’ll get into below, visibility increasingly depends on how content is interpreted and cited by LLMs, so distribution is no longer just a logistics decision, it’s a signal strategy.
Spend five minutes on LinkedIn and you’ll see every possible take on AI search:
“SEO is dead. “
“AEO is the future.”
“Just do good ol’ SEO.”
The truth is more nuanced. Buyer behaviour is shifting, models are reshaping discovery, and we’re all (whether we admit it or not) still figuring out what works. Anyone claiming to have it nailed is overselling.
Here’s how I think strategic marketers are responding…
Visibility Isn’t About Clicks Anymore, It’s About Citations
For the first time in 20 years, even Google is sounding the alarm. As users turn to LLM “answer engines” like ChatGPT, Gemini, and Perplexity, query volume is slipping, and some of that traffic, according to leaked internal memos, “is gone for good.”
Analysts at Gartner go further, forecasting a 25% drop in traditional search traffic by 2026 as conversational AI absorbs informational queries.
The shift is already underway: Semrush found that 13% of March 2025 searches already surfaced an AI Overview—double January’s share. One of the surprising advantages of this shift? Speed. Unlike traditional SEO where updates could take weeks to register, marketers are seeing LLM answers change within a day of updating core content. In some cases, models will pick up a fresh blog post before Google even crawls it. That makes your brand story more responsive than ever… and your old content liabilities more visible, too.
Visibility Now Lives Inside the Model
When a prospect asks Perplexity “Is Company X a reliable fintech vendor?”, the answer rarely comes from Company X’s website. Instead, the model stitches together snippets from SEC filings, podcast transcripts, Forbes articles, and even the careers page from your HR team. This reframes the press release debate on LinkedIn. The real question isn’t whether to issue a release, but whether every asset you push into the public domain is structured and credible enough to train a model in your favour.
Brand Mentions Are the New Backlinks
Ahrefs analyzed 75,000 domains and confirmed what savvy SEOs have started to suspect: the top predictor of appearing in Google’s AI Overview isn’t backlinks or Domain Rating, but unlinked brand mentions on authoritative sites. The correlation? ρ = 0.664—triple the strength of traditional backlink signals.
What reliably produces clean, machine-readable brand mentions at scale? Yes press releases, but also podcast transcripts, awards, analyst reports, syndicated updates, AKA the rest of a holistic marketing strategy.
Press Releases Aren’t Cool; They’re Useful
This isn’t an argument for blindly reviving old PR playbooks. Today, PR is less about racking up backlinks and more about shaping the story people, and search models, tell about you.
Press releases are helpful not because of the format itself, but because they generate structured, high-trust signals in places models pay attention to. They work when they’re part of a coordinated effort to build authority across multiple surfaces, not when they’re fired off as one-off blasts or “random acts of marketing.”
Releases give you three GEO advantages:
Structure - Headline, dateline, and boilerplate mirror the markup LLMs crave.
Distribution - Syndication across hundreds of news pages plants consistent brand mentions.
Authority - Newswire domains are seen as high-trust sources by both search engines and LLMs, giving your message more weight than content published on lightly-trafficked brand blogs.
Don’t blow your whole marketing budget on wire fees. Releases are one tool, not the strategy. Interview quotes, awards lists, data drops, and thought leadership do just as much to feed the model.
If someone asks ChatGPT, “What’s the best payroll software for startups?” it’s not just pulling from search rankings. It’s pulling from everything it’s seen—product comparisons, user reviews, podcast transcripts, even offhand mentions in blog posts or funding announcements. That’s the shift. You’re no longer optimizing for clicks, you’re optimizing to be part of the conversations.
Again, this is the part tactical distribution questions tend to miss. The teams getting ahead are thinking in four tracks at once:
Content – Moving from keywords to answers to personalized answers.
Technical – Shifting from on-page SEO tweaks to automated, well-structured, lightning-fast sites.
Authority – Evolving from link building to securing widespread, positive mentions.
Measurement – Swapping keyword rank reports for share of voice, sentiment, and AI Overview presence.
Strategic marketers aren’t just thinking about placement; they’re thinking about brand presence and architecting every external-facing asset to create clean, consistent signals that compound over time. Treating GEO as a maturity model, not a single tactic, keeps you from chasing every shiny object.
From SEO to GEO: The Playbook Has Flipped
| Yesterday (SEO) | Today (GEO) |
|---|---|
| Optimize for crawlers | Optimize for LLMs |
| Internal links, page speed | Third-party credibility, clear summaries |
| Keyword density | Novelty and “information gain” |
What this means in practice:
Think like a signal architect – The smartest teams recognize every external touchpoint (e.g., newswire hits, Glassdoor reviews, speaker bios, etc.) feeds the model and operate more like “brand signal architects,” ensuring messaging is consistent across channels.
Embed PR into SEO and content workflows – Align earned and owned content to reinforce the same signals. Press hits, blogs, and bylines should echo core narratives to boost citation relevance.
Build branded ecosystems, not one-offs – Communities, microsites, resource hubs, and founder-led content all help models associate your brand with a topic cluster. Authority isn’t just earned through media, it’s reinforced through repetition and thematic cohesion.
Bullet your value upfront - LLMs (and humans) skim. If you’re going to issue them, press releases that open with 3–5 crisp bullets see higher mentions in Semrush’s citation tracking.
Lead with unique information - LLMs prioritize “information gain”: unique data, clear POVs, and original framing.
Measure what matters now - Don’t just track backlinks and impressions. Monitor your presence in AI Overviews, share of voice across earned and owned channels, query-level sentiment, and the pipeline impact of organic content. That’s the new narrative strategic marketers are building.
Fewer Plays, Bigger Impact
Buying committee insights by Dentsu, TrustRadius, and Forrester reveal most B2B deals are won before a sales rep even enters the conversation, and that 78% of buyers choose vendors they’ve heard of before research begins.
Brand marketing isn’t a “nice-to-have” or some top-of-funnel vanity metric. It’s what gets you on the shortlist, lets you hold your price, and keeps customers around for the long haul. The quickest way to grow revenue isn’t cramming another tactic into Q3—it’s making your brand more memorable and more trusted.
Right now, LinkedIn is full of advice to start a newsletter, launch a podcast, build a community, jam voice-search keywords into every blog post, and pump out 30 short videos a month. But chasing every trend usually leaves you with a graveyard of half-finished projects. The real constraint isn’t ideas, it’s capacity; headcount growth is flat, forcing “do-more-with-less” decisions. According to a Content Marketing Institute report, 64% of B2B teams say their top challenge is “too many priorities, not enough resources.”
It’s the same story across the industry: headcount growth is flat, inbound requests are up, and the temptation is to keep stacking on tactics until you’re spread paper-thin. In the GEO era, that’s a losing game. The win comes from picking a few plays you can execute brilliantly and making them impossible to ignore, in both human and machine memory.
Here are the LLM-friendly brand marketing plays I’m prioritizing in Q3:
| Move | How it Pays for Itself (x3) | Supporting Data |
|---|---|---|
| Original research / Proprietary data |
• Earned media magnet • Sales enablement proof point • Evergreen SEO/GEO citation fuel |
Reports with original research and unique data generate more backlinks than opinion- and product-focused content (Cision 2024) |
| Short-form video cut from webinars and podcasts |
• Thought leadership snackables • Ad creative with top ROI • Social proof for future events |
Short-form video delivers the highest ROI for 1 in 5 marketers (HubSpot 2025) |
| Customer story roundtables |
• Event-driven pipeline driver • Podcast episode and blog series • Social proof snippets for sales decks |
58% of B2B marketing executives rely on their networks to build a shortlist (Wynter 2024) |
Final Thought: Prioritize What Makes You Worth Citing
Press releases aren’t a silver bullet. They never were. In AI search, they’re useful for a new reason: they create structured, high-trust signals that models can index and cite. When integrated into a broader content strategy, they help build visibility where it matters: inside the model.
This isn’t a call to revive legacy PR tactics for old times’ sake (I started in PR and I get the nostalgia). In 2025, your audience is both human and machine, and every external-facing asset is a training input. The question isn’t just where it gets published, it’s whether it helps the model trust you.
Brand marketing isn’t a soft play, it’s the reason you're in the room. Yet, most B2B teams are stretched thin, stuck in a cycle of fragmented execution. So the smarter question for Q3 isn’t “What else can we do?” It’s “What’s worth doing well?” GEO rewards the same fundamentals we’ve always known, but punishes half-baked execution faster. Do fewer things. Do them so well they get cited—by people, by publications, and by the models shaping the next era of search.
I’m betting on assets and activities that build salience, earn trust, and build long-term value, AKA content worth citing.
Anything else is noise.