A marketer clicks a highlighted brand name inside an AI answer… and a right‑side panel opens. It looks a lot like a Knowledge Panel. But it’s inside ChatGPT.

And the weird part?

The “story” in that panel isn’t pulled from the brand’s homepage. It’s stitched together from third‑party sources across the web. The sources are listed right there. You can see them.

That’s the game now.

AI trust is earned off‑site.
And if your brand isn’t being referenced off‑site, by the kinds of sources AI trusts, then you’re asking an algorithm to believe you based on… your own marketing.

Which is adorable.

This is a playbook for building earned authority that influences humans and AI: press, thought leadership, “as‑seen‑in,” press kits, spokesperson credibility, and the quiet-but-powerful concept behind all of it: Co‑citations.

What you’ll learn?

• Why off‑site signals matter more in the AI era (and why “we said so” doesn’t work anymore)?
• What co‑citation is (and how to build it without spam links)?
• How to turn PR into AI citations, intentionally, not accidentally?
• Which linkable assets actually work for GEO (and why benchmark reports are the cheat code)?
• How leaders should publish in the AI era (hint: “who said it” matters again)?

Why off‑site signals matter?

Let’s say you have the best content on the internet about your topic.

Clear. Helpful. Original.
Your writers cried. Your designer was mildly offended by your feedback.
You shipped it.

And then the AI answer cites… someone else.

That’s not always because your content is worse. It’s often because your brand isn’t the most trusted “source object” in the broader web graph.

AI answers are built on sources, not vibes

Modern “generative engines” don’t just generate from memory; they typically retrieve documents and then produce a response grounded in those sources, with attributions/citations so users can verify claims.

So if your brand wants to show up in AI answers, you’re not only competing on “ranking.” You’re competing on:

• Source eligibility (will the engine retrieve you?)
• Source trust (will it cite you?)
• Entity confidence (does it know who you are, and what you’re “about”?)

And if you’ve seen the “entity panel” behavior in AI, where the system tries to answer “what is this thing?” you’ve already felt the shift from keyword ranking to entity resolution.

Humans and AI both do the same thing: they look for third‑party proof

Google’s own guidance is blunt about trust.

Google says its systems use a mix of factors aligned with E‑E‑A‑T (Experience, Expertise, Authoritativeness, Trust), with trust as the most important.

And the Search Quality Evaluator Guidelines (the rater handbook) repeatedly emphasize a principle that applies perfectly to AI, too:

• Be skeptical of what websites say about themselves.
• Look for independent reputation information: reviews, expert references, news articles, and credible third‑party sources.

If you’re reading this as PR/Comms or a marketing leader, you already know this in your bones:

“We need third‑party proof.”

The only twist is that now you need third‑party proof that influences both humans and AI.

See Also: AI Visibility: How to Show Up in AI Search & Answers

What’s co‑citation (and why does it matter)?

Co‑citation is one of those terms that sounds like an SEO buzzword… until you see what it really means.

Search Engine Journal’s plain‑English definition:

Co‑citation means a website is mentioned by two different sources, but not necessarily linked.

That’s it.

Two independent sources mention you (or mention you alongside something else). The web graph learns an association. And over time, your entity becomes more “connected” to the concepts you want to own.

Co‑citation isn’t a link scheme. It’s a reputation pattern.

A link is a strong explicit signal. A mention is a weaker implicit signal.

But mentions scale in a way links often don’t, because journalists and creators mention brands constantly without linking. And AI systems that build responses from third‑party sources are swimming in mentions.

If AI is stitching answers from sources “across the web,” then being repeatedly mentioned across that web matters.

Co‑citation vs. co‑occurrence (quickly)

• Co‑citation: two sources mention the same entity (or mention entity A + entity B together).
• Co‑occurrence: terms/entities appear near each other in context (semantic proximity).

For our purposes, you can treat them as one strategic idea:

If your brand repeatedly appears alongside the right entities in the right sources, you become easier to categorize, retrieve, and cite.

The core narrative: AI trust is earned off‑site

Here’s the simplest model:

1. AI systems prefer credible sources (and often show citations so users can verify).
2. Credibility is reinforced by independent reputation information (not your own claims).
3. PR + thought leadership + linkable assets create independent mentions.
4. Those mentions create co‑citation patterns.
5. Co‑citation patterns increase the odds you’re included as a source and cited.

No hacks. No “AI trick.”
Just reputation. At scale.

Key #1: Why off‑site signals matter

Let’s make this painfully practical.

On‑site is what you say. Off‑site is what the world says.

Your website is your testimony.

PR, press, quotes, awards, conference lineups, partner pages, directory listings, independent reviews, and benchmarks?

That’s corroboration.

And Google’s rater guidelines explicitly tell raters to look for independent sources like expert recommendations, news articles, and other credible information when judging reputation.

AI systems aren’t “raters,” but the principle is the same:

Independent sources reduce uncertainty.

If your brand is invisible off‑site, AI has no reason to pick you

If you’re not mentioned, AI can’t “find you” as often in source retrieval.

If you’re mentioned but only on low‑credibility sites, you look risky. 

If you’re mentioned in credible places, with consistent descriptors, you become… safe to cite.

Key #2: Co‑citation strategy (safe + credible)

Let’s talk about building co‑citations without becoming ‘That’ Brand.

You know the one.  The one that just sounds a little too good to be true.

The safe way to build co‑citations

Think in three layers:

1. Entity clarity (who are you?)
2. Context clarity (what are you known for?)
3. Source credibility (who is saying it?)

Google’s documentation stresses “Who” created content, bylines, author pages, and background, because it helps people understand E‑E‑A‑T.

And the rater guidelines tell evaluators to seek independent reputation information, including expert recommendations and news.

So: if you want co‑citations that matter, you need clean entity signals and credible corroboration.

The Earned Authority Co‑Citation Map (framework)

Before you pitch a single journalist, do this mapping exercise.

Step 1: Pick your “authority neighbors”

These are the entities you want to be mentioned near:

• Standards (SOC 2, HIPAA, ISO 27001)
• Platforms (HubSpot, Salesforce, Shopify)
• Categories (“B2B demand gen agency,” “fractional CMO,” etc.)
• Competitors (yes, really)
• The “big concepts” in your space

Why competitors?

Because being mentioned in the same conversation can elevate you into the same mental bucket. (And co‑citation is literally the signal of “mentioned together.”)

Step 2: Identify the “sources AI trusts” for your category

Think:

• Industry publications
• Trade journals
• Major podcasts
• Research orgs / universities
• Professional associations
• Established SaaS review platforms
• Government / NGO sources (when relevant)

(Notice what’s not on the list: random blogs with 400 words and 30 affiliate links.)

Step 3: Write your canonical descriptor

This is the phrase you want repeated across the web.

Example format:

[Brand] is a [category] that helps [ICP] achieve [outcome] using [differentiator].

This is not tagline work.
This is entity resolution work.

Because AI systems try to answer “what is this thing?” and build summaries from what they trust.

Step 4: Choose 3–5 proof points that are cite‑worthy

Not fluff. Not “innovative.”

Proof points that a journalist would actually include:

• A benchmark number (from your research)
• A quantified case study result
• A quote from a recognizable expert
• An award, certification, or credible partner

And yes, GEO research suggests that including citations, quotations, and statistics can increase visibility in generative responses.

So you don’t just want PR coverage.
You want PR coverage that contains cite‑worthy artifacts.

See Also: AI Signal Map: What AI Uses to Read & Trust Your Site

Key #3: As‑seen‑in + press kit (the AI‑friendly version)

Most “press kits” are built like a folder dump.

Logos. Headshots. A PDF from 2019.
“Please don’t use the old logo.”
(You used the old logo.)

Instead: build your press kit as an authority hub.

What to include in a modern press kit?

You should include components like:

• Leadership bios
• Company history / timeline
• Photo library
• Logos (with proper formats and guidelines)
• Case studies and testimonials

That’s the minimum.

Now add the AI era layer:

1) A canonical “About” paragraph (copy/paste ready)

This becomes your consistent descriptor that travels.

2) A spokesperson “fast facts” box

• Full name
• Title
• Topic areas
• Short credential list
• Prior press/speaking highlights (even if modest)

3) A “what we believe” set of 3–5 statements

Why? Because journalists quote belief and POV.

4) Your linkable asset (your benchmark report, framework, or dataset)

Journalists love citing research.
AI loves citing research even more.

The “As Seen In” section isn’t decoration

Your as‑seen‑in bar can be a trust trigger for humans… but it can also be a structured “receipt” for crawlers if you implement it cleanly.

Make it real:

• Use actual publication names
• Link to actual coverage (where permissible)
• Keep it current
• Avoid fake logos, “featured on” abuse, or paid placements disguised as earned

Remember the rater guideline principle: be skeptical of self‑promotional claims; independent sources matter.

You’re not trying to impress your mom.
You’re trying to reduce uncertainty for algorithms.

Key # 4: Linkable assets (benchmark report)

If you want consistent earned mentions, you need something worth mentioning.

That’s what linkable assets are: conversation fuel.

And in the GEO research, the authors explicitly show improved visibility when content includes citations, quotes, and statistics.

So let’s build the kind of asset that naturally contains those.

The benchmark report (why it works)

A benchmark report does three things at once:

1. Creates original, cite‑worthy numbers
2. Gives journalists a reason to cover you
3. Forces your brand to be mentioned in context with key entities and terms

It’s a co‑citation generator.

What a “GEO‑friendly” benchmark report looks like

Not “State of Marketing 2026” with recycled stats.

A useful benchmark report is:

• Narrow (one specific behavior/outcome)
• Comparable (clear methodology + segments)
• Quote‑friendly (strong takeaway lines)
• Source‑rich (citations to what you’re building on)
• Reusable (charts, embed code, and clear copy/paste snippets)

Because if the web is going to cite you, you need to make citing you easy.

Examples of benchmark angles that earn citations

Pick one:

• “Median CPL by channel for [industry]”
• “Time to SQL by source (2026 benchmarks)”
• “Homepage conversion rates by traffic temperature”
• “Top objections in [industry] sales calls (analyzed from X calls)”
• “AI visibility scorecards: which sources get cited?”

And yes, you can do this without enterprise budgets.
You just need a good dataset and a clear narrative.

Benchmark report checklist (the stuff most teams forget)

• A methods section (sample size, time range, what was included/excluded)
• A “what this means” section (interpretation)
• A “how to use this” section (practical application)
• A downloadable version (PDF) AND a web version
• A press kit tie‑in (pre‑written bullet takeaways + charts)

Also: include quotable statements and statistics prominently. That aligns with the GEO findings about visibility lifts from citations/quotes/stats.

Key #5: Spokesperson authority (speaker/author)

This is where most brands underinvest.

They try to build “brand authority” while their leadership team is invisible or inconsistent online.

But in the AI era, authorship matters again.

Google explicitly recommends making it clear who created content, bylines, author pages, and background, because it helps people understand E‑E‑A‑T.

So the question becomes:

How should leaders publish in the AI era?

The old model: publish on your blog.
The new model: publish where authority already exists.

The spokesperson ladder

Start here and climb:

1. Company blog (with real bylines + robust author pages)
2. Guest contributions to respected industry sites
3. Podcast guesting (topic‑specific shows in your niche)
4. Conference speaking (even small stages count if recorded/posted)
5. Quoted expert (HARO‑style, journalist requests, PR relationships)
6. Original research author (benchmark report)
7. Book / long‑form cornerstone (if appropriate)

Every rung produces more independent mentions, more co‑citations, more entity confidence.

The “author page” isn’t a bio. It’s a trust artifact.

A good author page includes:

• Name, role, and scope of expertise
• Credentials that matter for the topic
• Links to external profiles (LinkedIn, speaking reels, prior interviews)
• A list of authored content (so systems can cluster expertise)

Why this matters: it reduces the “who is this?” uncertainty for both humans and machines.

See Also: Trust Signals That AI Can Actually Use: Reviews, Certifications, and Content Integrity

How to turn PR into AI citations (the playbook)

This is the ICP question hiding under all the other ones.

So here’s the playbook.

Step 1: Stop pitching “announcements.” Start pitching “sources.”

AI citations don’t come from announcements.

They come from sources that answer questions.

So before you pitch PR, ask:

• What question does this coverage help answer?
• What claim can a writer cite?
• What asset will they reference?

Step 2: Build “citation hooks” into PR

A citation hook is anything that makes a journalist’s job easier:

• A benchmark statistic
• A defensible methodology
• A quote with a strong point of view
• A simple framework with memorable names

And again, the GEO paper’s finding that citations/quotes/stats improve visibility is a giant clue about what these engines reward.

Step 3: Ensure your brand is described consistently

If one article calls you a “PR firm,” another calls you a “content agency,” and another calls you a “GEO consultant,” you’re fragmenting your entity story.

Pick your descriptor.
Use it everywhere.

Step 4: Make the press kit the canonical hub

When someone wants to verify:

• Who you are
• What you do
• Who leads it
• What proof exists

…they should land on one page that makes it easy.

Step 5: Convert coverage into durable authority

PR is a spark. Earned authority is a system.

When you earn coverage:

• Add it to the As‑Seen‑In page
• Add it to the spokesperson page (“Featured in…”)
• Add it to the press kit
• Use it in sales enablement
• Use it in investor decks
• Use it in recruiting

Why? Because reputation is cumulative.

And independent third‑party references are exactly what quality evaluation frameworks instruct people to look for.

How to build authority without spam links?

Don’t chase links. Chase independent validation.

The Google rater guidelines literally call out that websites love telling users how great they are, and raters should not accept those claims without independent verification.

So the non‑spam path looks like:

• Earned media (real coverage)
• Partner pages (legitimate partnerships)
• Professional associations (real memberships)
• Conference speaker listings
• Expert roundups (with editorial standards)
• Original research that people cite
• Reviews (where appropriate)
• Community contributions (open‑source, templates, calculators)

The moment you try to “manufacture reputation,” you lose it

If the coverage is fake, paid, or irrelevant, it’s not earned authority.

It’s confetti.

What linkable assets work for GEO (besides benchmark reports)?

Benchmark reports are king, but they’re not the only option.

Here’s the shortlist of assets that naturally earn citations:

1. Original research (survey, dataset analysis, benchmarks)
2. Industry glossary / definitions (when done at expert depth)
3. Calculators / estimators (ROI, budget, forecasting)
4. Templates (audit checklists, PR pitch templates, playbooks)
5. Interactive tools (scorecards, graders, validators)
6. Open datasets (even small ones, if useful)
7. Case study libraries (with real numbers)
8. Visual frameworks (easy to reference and explain)

The key: make it cite‑worthy.

If your asset doesn’t contain anything a third party can reuse, quote, or reference… it won’t travel.

The 90‑day Earned Authority plan (PR + co‑citations)

If you need something operational, here you go.

Days
1–14: Build the foundation

• Define your canonical descriptor (one sentence)
• Build/refresh your press kit (bios, logos, photos, case studies, testimonials)
• Build/refresh your As‑Seen‑In page
• Create/refresh author pages and bylines (make “who” obvious)
• Create a list of 25 target sources (outlets/podcasts/associations) </p>

15–45: Build the linkable asset

• Pick the benchmark topic
• Collect data / run survey / compile analysis
• Write the report (with methods + charts)
• Create a press‑ready summary
• Add citations, quotes, and statistics intentionally (visibility matters)

46–90: Distribute and earn co‑citations

• Pitch the asset to journalists and editors
• Offer exclusives to top targets
• Book podcasts / webinars around the findings
• Publish 3–5 thought leadership pieces tied to the data
• Repurpose into short “insight posts” for LinkedIn (for humans)
• Track mentions and update your authority hub

Measurement: how do you know this is working?

You’re not measuring “traffic.” You’re measuring trust signals.

Track:

• Number of independent mentions (press, blogs, podcasts, listings)
• Co‑citation proximity (are you mentioned alongside your target “neighbors”?)
• Spokesperson mentions (is the leader being cited by name + title?)
• AI citations (does your brand show up as a source in AI answers for your target queries?)
• Brand SERP changes (more third‑party results, panels, and consolidated entity signals)

And if you’ve watched the entity panel behavior in AI, you already know the north star:

Can the system confidently summarize who you are, using third‑party sources?

FAQ 

How do we turn PR into AI citations?

You design PR as source creation:

• publish a cite‑worthy asset (benchmark report)
• build a press kit hub
• pitch credible outlets
• ensure consistent descriptors and proof points
• accumulate independent mentions

GEO research supports that citations/quotes/stats can increase visibility in generative responses.

What’s co‑citation and why does it matter?

Co‑citation is when your website/brand is mentioned by two different sources, even without links.
In an AI world built on third‑party sources, being repeatedly referenced in credible contexts increases the odds you’re retrieved and cited.

What linkable assets work for GEO?

Original research and benchmark reports are the most reliable, because they generate statistics, quotes, and citations, exactly the ingredients shown to improve visibility in GEO research.

How should leaders publish in the AI era?

Make authorship obvious and consistent (bylines + author pages) and publish off‑site where authority already exists. Google explicitly encourages accurate authorship info and bylines where expected.

How do we build authority without spam links?

Focus on independent validation: earned media, expert references, original research, reputable associations, real speaking engagements. The rater guidelines emphasize independent reviews and credible sources over self‑claims.

Bottom line

If your strategy is “publish more on our blog,” you’re playing single‑player mode.

The AI era is multiplayer.

You win by earning authority off‑site, through press, thought leadership, and linkable assets that generate co‑citation signals.

Because AI trust isn’t claimed.

It’s corroborated.

See Also: AI Visibility Scorecards & Reporting | How to Measure AI Search Presence

About The Author

Dave Burnett

I help people make more money online.

Over the years I’ve had lots of fun working with thousands of brands and helping them distribute millions of promotional products and implement multinational rewards and incentive programs.

Now I’m helping great marketers turn their products and services into sustainable online businesses.

How can I help you?

See author's posts