AI Visibility

How to Optimize Earned Media for GEO: 7 Tactics to Improve AI Visibility (2026)

Seven tactics for pitching useful earned media coverage and measuring whether it appears in AI answers.

Jaxon Parrott
Jaxon ParrottJan 26, 2026

To optimize earned media for GEO, start with the buyer questions and publications AI answers already cite; pitch a verifiable finding or expert perspective to those publications, then record whether the resulting coverage is cited in repeat tests. A placement is evidence to test, not a promise of a citation. Muck Rack's May 2026 What Is AI Reading? analysis of more than 25 million links in ChatGPT, Claude and Gemini responses across 17 industries found 84% came from earned media, the sources brands neither own nor pay for — journalism, third-party corporate content, Wikipedia, Reddit, government, and academic sources together, with journalism alone at about 27%. Editorial coverage is the slice of that pool PR competes for. Clear, verifiable evidence makes coverage useful to readers and gives engines material they may cite. This seven-step playbook covers how to optimize earned media for GEO — from targeting relevant publications to measuring citations over time. Every step is built on Machine Relations principles: the discipline of making your brand legible, retrievable, and credible inside AI-driven discovery systems.

How to optimize earned media for GEO: a practical test

  1. Baseline the question. Record the exact buyer question, engine, date, answer, named brands and cited URLs. Repeat the same question over time; one answer is a snapshot, not a ranking.
  2. Choose a coverage target. Compare the publications cited for that question with relevant editorial outlets in your category. Check their recent coverage and pitch a genuinely useful, independently verifiable finding or expert perspective.
  3. Prepare the evidence. Give the journalist the source, method, date, context and a clear description of the brand. A journalist controls the published story; a pitch cannot guarantee wording, a link or placement.
  4. Measure separately. After publication, log the article URL and publication date. On later checks of the same question, record a brand mention and a link to that specific article as separate outcomes. A citation of an outlet does not by itself prove that your placement caused an answer or recommendation.

For example, a cybersecurity team can record the sources cited for “which tools help small businesses implement zero trust?”, pitch a documented security finding to a relevant trade publication, and then check whether later answers mention the brand and cite the resulting article. If the outlet is cited but that article is not, the placement has not been shown to be the cited source. Use those observations to choose the next pitch, not as proof that coverage caused a change.

Key Takeaways

  • Journalism accounts for about 27% of cited links, inside an earned-media band (non-owned, non-paid) of 84% — Muck Rack's May 2026 What Is AI Reading? edition puts editorial coverage alongside third-party corporate content, Wikipedia, Reddit, government, and academic sources in the links ChatGPT, Claude, and Gemini cited across 17 industries.
  • Specific, sourced data points lifted visibility by up to 40% in the GEO paper's benchmark — Aggarwal et al. (KDD 2024) measured the effect of adding statistics, quotations and cited sources on their own GEO-bench test engine; it is a benchmark result, not a live citation rate.
  • Related coverage merits testing — Stacker and Scrunch reported a 239% median visibility lift in an observational comparison of publisher-network and owned-domain citations for 87 distributed stories from 30 brands across eight platforms.
  • Source mixes change over time. Muck Rack's October 2026 What Is AI Reading? edition found that more than 84% of cited links came from earned and third-party sources, meaning sources brands neither own nor pay for (journalism, government and NGO websites, academic research, encyclopedic sources and third-party corporate content). In September owned media was 15.2% of citations and paid and advertorial content 0.3%, and journalism alone was 22%, down from 27% in May. The September analysis examined prompts across ChatGPT, Claude, Gemini and Perplexity.
  • GEO is one layer inside Machine Relations — GEO optimizes how content reaches AI engines (Layer 4: Distribution). The full Machine Relations system — authority, entity, citation, distribution, measurement — is what makes individual GEO tactics compound.

What Muck Rack’s sample found

In Muck Rack’s May 2026 sample, 84% of cited links came from earned media, sources brands neither own nor pay for. Its analysis covered links in ChatGPT, Claude and Gemini responses across 17 industries. That broad category bundles journalism with third-party corporate content, Wikipedia, Reddit, government and academic sources; journalism alone accounted for about 27%. Seer Interactive’s SearchGPT analysis reported that 87% of citations matched Bing’s top organic results.

This is a structural shift, not a trend. Traditional SEO optimizes your own website to rank. GEO optimizes your presence in the sources AI engines already trust. The mechanism is earned media — the same mechanism that built brand credibility with human readers for decades — now repurposed as the citation fuel for AI answers. BuzzStream and Citation Labs analyzed 3,600 AI prompts across 10 industries and found 81% of AI news citations come from original editorial content — while press releases accounted for just 0.21%.

Gartner predicted in February 2024 that traditional search engine volume would drop 25% by 2026 as AI-driven discovery expanded. G2's survey of 1,000+ B2B buyers (August 2025) found that 87% say AI chatbots are changing how they research software, and 50% now start their buying journey in an AI tool. An owned page can be cited too; independent coverage gives buyers and engines additional sources to consult, but does not guarantee a citation or recommendation. Whether earned media improves AI search results at all, and which strategies win, is set out in How Earned Media Now Dominates AI Search Results.

AuthorityTech's eight years of results-based PR and 1,500+ direct editorial relationships inform its approach to this work. As founder Jaxon Parrott puts it: "PR built authority with human readers through editorial relationships and earned media. Machine Relations builds authority with machine readers through the same relationships and the same earned media." For a deeper look at the timeline, see the earned media AI citation timeline.

How to Identify Which Publications AI Engines Actually Cite

Not all publications carry equal weight in AI citation systems, and which ones matter is measurable. In the Machine Relations Index release of October 17, 2026, editorial publications were the largest role-classified source on every one of the six buyer question shapes the Index tracks, from 24.7% of classified citations on problem-first questions to 34.9% on best-of questions, while wire and press-release distribution took 0.8%. The publications themselves change with the category: in the October 26, 2026 release, the most-cited editorial publication in enterprise software best-tools answers was TechRadar, in 24 of 208 answer runs, while consumer finance best-tools answers cited Forbes in 82 of 249.

A publication engines already cite for your buyer’s question is a potential coverage target. Measure whether the specific placement URL later appears in answers; an outlet citation alone does not establish that it does. AuthorityTech's publication pages show each publication's standing by category and question shape, and how to get into each one. How to split a GEO program between owned and earned media is a separate decision, set out in our GEO earned media strategy framework.

To map which publications AI engines cite in your category:

  1. Ask ChatGPT, Perplexity, and Gemini the exact questions your target buyers ask. Note every source citation.
  2. Cross-reference cited publications against your existing media targets. Prioritize overlaps.
  3. Track citation frequency over time — AI engines update their source preferences as publications gain or lose editorial authority.

AuthorityTech maintains a running audit of which publications appear in AI citations across categories. The pattern is clear: publications with strong editorial standards, named bylines, and original reporting dominate AI citations. Content farms and pay-to-play outlets rarely appear.

How to Optimize Earned Media Headlines for AI Query Matching

Make the proposed story clear and relevant to the questions buyers ask. A fundraising announcement and a problem-solving article serve different purposes. Suggest an angle that accurately connects the company’s evidence with the reader’s question; an editor decides the final headline and wording, and a particular headline does not guarantee a match or citation.

The optimization process:

  1. Research query phrasing. Use Google Search Console, Ahrefs, or ChatGPT search prompt data to understand how your audience phrases questions.
  2. Pitch query-aligned headlines. Before pitching a story, draft 2–3 headline variants that match buyer queries. Share these with editors as suggested angles.
  3. Offer useful context. Provide clear supporting facts and possible subtopics; journalists control headlines, subheadings and the final article.

This is where GEO overlaps with the broader Machine Relations stack. Layer 4 (Distribution) includes both GEO and AEO — tactics that work best when the underlying earned media placement is already query-aligned. For an argument about the role of independent coverage, see why earned media matters to GEO.

How to Structure Data Points for Maximum AI Extraction

Specific, attributed data points give a generative engine more to surface than vague claims do. The GEO paper (Aggarwal et al., KDD 2024) found, on its GEO-bench benchmark, that adding statistics, quotations or cited sources lifted a position-adjusted visibility score by 30–40% for its top methods, with the effect varying by domain. That is the authors' own test engine rather than a live citation rate, but it is the controlled evidence behind the practice. Ahrefs' research across 75,000 brands adds another dimension — brand web mentions correlate more strongly with AI Overview visibility than backlinks (0.664 vs 0.218 correlation).

When relevant and verifiable, provide evidence an editor can assess, such as:

Data TypeIllustrative Example
Percentages and growth metrics"Reduced costs by 47%" with source and time period
Comparative claims"3x faster" with the comparison and method
Industry benchmarksAn industry average with its source and comparable population
Named methodologies"Using the Machine Relations framework" with a description of the method

Give journalists data that is accurate, attributed and useful in context; tables can make comparisons easier for readers to inspect, but format alone does not guarantee an AI citation. AuthorityTech’s approach to measuring earned media ROI for AI visibility tracks whether particular coverage appears in answers.

Should Original Research Be Used as Earned Media in GEO?

Yes: original research can give an editor a new finding and give readers a sourced number to check. On the GEO paper's benchmark, adding statistics, quotations and cited sources lifted visibility by 30–40% for its top methods, a benchmark result rather than a live citation rate, and a finding only you can publish is the purest form of that intervention.

Run it in three moves. Publish the full dataset and method on your own site, so editors and engines can check the number at its source. Pitch the finding, not the company, to the publications engines already cite for the question it answers. Then keep the figure, its date and its denominator identical in every placement and on your own page, so the coverage and the source corroborate each other instead of drifting apart.

How to Build Entity Consistency Across Earned Media

Accurate, consistent company facts make independent coverage easier for readers to understand and verify. Supply the correct name, product description and current titles in press materials, and flag factual errors through the publication’s correction process. Editors may describe a company in their own words; there is no evidence here that wording differences alone prevent a citation.

The entity consistency audit checklist:

  • Company name: Exact spelling and capitalization across all coverage (e.g., "AuthorityTech" not "Authority Tech" or "authority-tech")
  • Founder/executive names and titles: Consistent formatting with correct current titles
  • Product descriptions: An accurate description, adapted to the story without changing the facts
  • Category positioning: Clear category language appropriate to the publication and audience

Jaxon Parrott coined Machine Relations in 2024 to name the discipline of building authority in AI-mediated discovery. A clear description helps people understand the category; independent publications choose how they describe it.

How to Test Related Coverage Across Publications

Related coverage can give buyers and AI systems more than one independent source to consult. Choose distinct, relevant angles and measure the resulting answers; there is no established number of placements that guarantees a durable citation or a default recommendation.

An illustrative pitch cluster for a cybersecurity company targeting "zero-trust architecture" queries (outlet names and headlines are examples, not claimed placements):

  1. "How Zero-Trust Architecture Prevents Breaches" — TechCrunch (problem framing)
  2. "The Real Cost of Data Breaches in 2026" — Forbes (data evidence)
  3. "Why SMBs Are the New Target for Ransomware" — Inc (audience expansion)
  4. Expert quote in broader industry roundup — Wall Street Journal (authority signal)

Different angles give editors and readers different reasons to engage with the same area of expertise. Stacker and Scrunch reported a 239% median visibility lift in an observational comparison of publisher-network and owned-domain citations for 87 distributed stories from 30 brands across eight platforms.

Track whether individual articles appear in repeated answers; related coverage may provide additional sources, but the results do not by themselves prove a multiplier. Read more about the citation economy.

How to Maintain Recency Signals for AI Citation Priority

Keep facts and coverage current when the buyer’s question depends on current information. Review older citations for changed product details, outdated claims or broken source links. There is no supported universal expiry date for an earned-media citation; check the actual answers before deciding what needs a fresh pitch.

An illustrative planning cadence for earned media:

Publication TierPlanning cadenceExamples
Tier 1 (major business/tech)1–2 placements per quarterForbes, TechCrunch, Business Insider, VentureBeat
Tier 2 (industry trades)2–4 placements per monthSearch Engine Land, Adweek, industry verticals
Ongoing expert commentaryContinuousQuotes in relevant stories, roundups, analyst briefings

Treat this cadence as an editorial planning example, not a required threshold or guarantee. Adapt it to genuinely newsworthy material and the results of repeated checks. Together, coverage and measurement form the earned media citation infrastructure that Machine Relations is built on.

How to Monitor and Iterate AI Citations from Earned Media

AI citation monitoring is the feedback loop that turns earned media from a one-time placement into a compounding visibility system. Without measurement, you cannot know which placements are being cited, which queries trigger your brand, or where competitors are appearing instead of you.

The monitoring process:

  1. Weekly citation audit: Repeat the same unbranded buyer questions under the recorded engine and settings. Record brand mentions and exact cited URLs separately; use branded questions as a separate diagnostic.
  2. Source attribution tracking: Count which exact placement URLs appear in the observed answers and how often, alongside the number of tests. A brand mention is not necessarily a citation, and a cited placement does not by itself establish causation.
  3. Gap identification: Queries where competitors appear but you don't are earned media targets. Prioritize the queries with highest buyer intent.
  4. Iteration: Compare patterns across repeated checks and refine publications and angles. Do not assume a placement caused a change without a controlled comparison.

Tools like BrightEdge, Semrush, and specialized AI visibility dashboards can automate parts of this — but they diagnose the gap without closing it. AuthorityTech's GEO audit framework combines citation monitoring with the earned media execution needed to act on what the data reveals. The measurement without the mechanism is just a dashboard. The mechanism without the measurement is just hope.

If you would rather hire the whole loop than run it, AuthorityTech, The World's First Machine Relations Agency, chooses publications by what engines cite for a client's category, earns the coverage and judges it in the answers. Our ranking of the best earned media and AI PR agencies for AI search compares it with the rest of the field.

Where GEO, AEO, and SEO Fit Inside Machine Relations

GEO, AEO, SEO, and Digital PR are not competing alternatives — they are different layers of the same system. Machine Relations, which Jaxon Parrott coined in 2024, is the full architecture that contains each of them:

DisciplineOptimizes forSuccess conditionScope
SEORanking algorithmsTop 10 position on SERPTechnical + content
GEOGenerative AI enginesCited in AI-generated answersContent formatting + distribution
AEOAnswer boxes / featured snippetsSelected as the direct answerStructured content
Digital PRHuman journalists/editorsMedia placementOutreach + storytelling
Machine RelationsAI-mediated discovery systemsResolved and cited across AI enginesFull system: authority → entity → citation → distribution → measurement

GEO and AEO are tactics within Layer 4 (Distribution) of the Machine Relations stack. GEO can support both owned and earned content. An owned page can earn AI citations without earned coverage; relevant independent coverage adds another potential source rather than guaranteeing an outcome. The steps above help teams plan coverage and evaluate the answers that follow.

Frequently Asked Questions

Should original research be used as earned media in GEO?

Yes. Original research gives editors a reason to cover you and gives AI engines a sourced figure to cite, and the GEO paper's benchmark found that adding statistics, quotations and cited sources raised visibility by 30–40% for its top methods. Publish the method on your own site, pitch the finding to the publications engines cite for the question, and keep the number identical everywhere it appears.

Who coined Machine Relations and what does it mean for GEO?

Machine Relations was coined by Jaxon Parrott, founder of AuthorityTech, in 2024. It is the discipline of building a brand’s visibility and authority in AI-mediated discovery, including whether AI systems cite, surface and recommend it. GEO is one tactic within Layer 4 (Distribution) of the Machine Relations stack — the layer that handles how optimized content reaches AI engines. Machine Relations provides the full system that makes GEO effective: authority building, entity resolution, citation infrastructure, distribution, and measurement.

What did Muck Rack find about earned media and AI citations?

In Muck Rack’s May 2026 sample of ChatGPT, Claude and Gemini responses across 17 industries, 84% of cited links fell in its earned-media band: sources brands neither own nor pay for, including journalism, third-party corporate content, Wikipedia, Reddit, government and academic sources. Journalism alone accounted for about 27%. This is a sample result.

How many earned media placements does it take to build AI citation authority?

There is no reliable universal number. One placement may be cited, while several may not be; the question, outlet, article and engine all matter. Track individual placement URLs and repeated answers, then adjust your outreach based on observed results rather than a promised threshold.

Is Machine Relations just SEO rebranded?

No. SEO optimizes for ranking algorithms — the goal is a top-10 position on a search engine results page. Machine Relations optimizes for AI-mediated discovery systems — the goal is being cited and recommended in AI-generated answers across ChatGPT, Perplexity, Gemini, and AI agents. The mechanism is different (earned media authority vs. on-page signals), the measurement is different (citation share vs. ranking position), and the scope is different (full system vs. technical content optimization). Moz's 2026 analysis of 40,000 queries found that 88% of Google AI Mode citations are not in the organic top 10 for the same query. The figure is the overlap between those two result sets.

How do AI search engines decide which sources to cite?

Source selection differs by engine and question. The GEO paper (Aggarwal et al., KDD 2024) found that adding statistics, quotations and cited sources improved a visibility measure on its own benchmark. Prepare accurate, attributable evidence, then test what appears in repeated answers.