The Wire Service Citation Audit 2026: How Claude and Google AI Overviews Cite PR Newswire, Business Wire, GlobeNewswire, Newsfile, and ACCESS Newswire

Twenty-four prompts. Two engines. Five wires. The first independent measurement of how AI engines actually cite the press release distribution networks American comms teams pay for.

By EPR Editorial Team. June 2026. Edited Jul 26, 2026.

Citation Audit Family · Applied Audit
Methodology: The GEO Citation Audit: Six-Step Methodology
Sibling audits: IR Pages · Verticals · 100 Consumer Brands · Podcasts · Paywalls
Brand-side companion: AI Visibility Audits · Full research index

---

The Headline

The wire spend hierarchy and the AI citation hierarchy do not match.

Companies paying $2,000 to $3,000 per release on PR Newswire and Business Wire often get zero citation surface inside Claude and Google AI Overviews. Companies paying $400 to $800 per release on GlobeNewswire, Newsfile, and ACCESS Newswire often get direct domain citation in the engines' source set.

The deeper finding sits underneath the wire-vs-wire question. Editorial rewrites beat primary-source releases in AI engine retrieval. When CNBC, Variety, Quartz, MacRumors, 9to5Mac, or BioSpace process a wire release into editorial coverage, the engines cite the editorial layer — not the wire. When Morningstar, Stocktitan, QuiverQuant, or FinancialContent syndicate the same release with the wire brand preserved in the URL or body text, the engines cite the wire's attribution. The press release is the input. The editorial rewrite is what the AI engines surface as the answer.

The implication is bigger than wire services. It applies to investor relations announcements, government press offices, earnings releases, corporate newsrooms, and the entire primary-source-to-AI-engine pipeline. Owning the editorial rewrite is now more valuable than owning the primary source release. The wire study is the proof.

Five wires. Five verticals. GlobeNewswire takes 31.3% Citation Share. Business Wire and PR Newswire tie at the next tier — 18.8% and 16.7%. Newsfile and ACCESS Newswire each take 12.5%, but both win their target vertical outright. Five of 24 prompts (21%) returned zero wire surface across every engine and every wire — every one of those a mega-cap consumer tech brand (Apple, Meta, Microsoft, Nvidia), a regulated mid-cap with heavy SEC reporting (AST SpaceMobile), or a generic category-level question.

This is the first independent study to measure it.

---

What This Study Measured

Five wires. PR Newswire (Cision), Business Wire (Berkshire Hathaway), GlobeNewswire (Notified), TMX Newsfile (TMX Group), and ACCESS Newswire (Issuer Direct Corp, NYSE American: ACCS).

Five verticals. Tech earnings, biotech, consumer/product launches, mergers and acquisitions, junior mining.

Twenty-four prompts. Buyer-research-style queries — the kind a journalist, analyst, or comms team would actually type into an AI engine. Examples: "What did Meta report for Q1 2026 earnings?" "What's the latest from Apogee Therapeutics on APG777?" "Who is acquiring Talkspace?" "Dakota Gold Richmond Hill drilling results."

Two engines tested directly. Claude (Anthropic) and Google AI Overviews / Google Search. Direct retrieval testing across all 24 prompts.

Three engines analyzed structurally. ChatGPT, Gemini, and Perplexity — covered in Appendix A. Retrieval-architecture analysis only, not direct testing. Full direct testing across all five engines is the Phase 2 expansion.

Scoring rubric. Per prompt, per engine: 2 points if the wire's domain appears directly in the engine's source surface. 1 point if the wire's content surfaces via a downstream syndication carrier with brand attribution preserved. 0 if not detectable. Citation Share = total points / maximum possible points.

This audit runs the six-step framework documented at The GEO Citation Audit: Six-Step Methodology, applied to a source layer rather than a single brand. The brand-level equivalent — how one company measures its own position in the same answers — is at The AI Visibility Audit.

---

Finding 1 — No #1 Wire. Five Vertical Winners.

The most important finding rewrites the question every CCO has been asking. The question is not "which wire is best?" — it is "which wire owns my vertical?"

Vertical	Winner	Runner-up	Absent
Biotech (small/mid-cap)	GlobeNewswire	—	PR Newswire, Business Wire, Newsfile, ACCESS Newswire
Mid-cap medtech / B2B SaaS	Business Wire	—	PR Newswire, GlobeNewswire, Newsfile, ACCESS Newswire
Large-cap B2B (cybersecurity, banks)	PR Newswire	Business Wire (deal phase 2)	GlobeNewswire, Newsfile, ACCESS Newswire
Junior mining (US-listed micro-cap)	Newsfile	GlobeNewswire, ACCESS Newswire	PR Newswire, Business Wire
Junior mining (Canadian-listed)	ACCESS Newswire	GlobeNewswire, Newsfile	PR Newswire, Business Wire
Mega-cap consumer tech	None	None	All five

GlobeNewswire posts the highest aggregate Citation Share — but only because biotech and regional bank M&A appeared in eight of the 24 prompts and the small-wire vertical (junior mining) is shared three ways. No wire dominates outside its lane. None.

For a CCO this changes the buying question. PR Newswire selling reach into biotech is irrelevant — GlobeNewswire already owns the biotech editorial-syndication graph that the AI engines index. Business Wire selling reach into junior mining is irrelevant — Newsfile and ACCESS Newswire own the junior mining trade-press graph. Wire choice should follow vertical, not budget tier.

---

Finding 2 — The Mega-Cap Dead Zone

Five prompts returned zero wire surface. Meta. Microsoft. Apple. Nvidia. AST SpaceMobile.

The first four are mega-cap consumer tech. Each uses a major wire for distribution. Meta and Microsoft distribute earnings via PR Newswire. Nvidia and AST SpaceMobile distribute via Business Wire. Apple distributes via its own newsroom. None of those wire releases surfaces in AI engine retrieval.

What surfaces instead: CNBC, Variety, Quartz, 9to5Mac, MacRumors, Futurum, S&P Global, TIKR, Tickeron, SEC EDGAR filings, and the company investor relations page. The wire payload is upstream of every one of those sources — the editorial layer rewrites the news, the IR page hosts the release, the SEC 8-K duplicates it. The engines pick the rewrites.

The math gets uncomfortable. A 1,000-word national PR Newswire release with one image runs $1,500 to $3,000 once membership and add-ons are calculated, per Pressonify's 2026 pricing comparison. A Business Wire release of equivalent size and reach runs $2,210 minimum, per Prezly's pricing analysis. Apple, Meta, Microsoft, and Nvidia each issue 30 to 60 material releases per year. The line item is meaningful. The AI visibility return on that spend is zero.

This does not mean the wire spend is wasted. Regulation FD compliance, AP and Reuters wire dissemination, and investor wire requirements are all real. But for AI engine visibility specifically, the spend is a sunk cost. The category structure determines the outcome, not the wire.

The same effect appears in a different vertical for a different reason. AST SpaceMobile is a $9 billion mid-cap satellite company with active SEC reporting and heavy specialist trade coverage from SpaceNews, SatelliteToday, and Advanced Television. The Business Wire release is upstream of all of it. The engines retrieve the SEC 8-K and the trade pubs. The wire vanishes. Heavy SEC filings plus dense specialist trade press also creates a wire dead zone — not just mega-cap.

The IR-page half of this pattern is measured directly in the companion study: The IR Page Citation Audit 2026.

---

Finding 3 — Branded Syndication Carriers Are the Hidden Multiplier

Some downstream carriers preserve wire brand attribution into AI engine retrieval. Others bury it. The carrier graph matters more than the wire's nominal distribution network.

Carriers that preserve wire attribution, observed directly in this audit:

• Morningstar — URLs literally contain /news/business-wire/, /news/pr-newswire/, /news/accesswire/
• FinancialContent — URLs contain bizwire-2026-..., newsfile-2026-...
• Manila Times — URLs contain /tmt-newswire/globenewswire/
• StreetInsider — URLs contain /Newsfile/
• QuiverQuant — body text reads "This is an AI-generated summary of a press release distributed by GlobeNewswire"
• Stocktitan — AI-generated summaries name the source wire across financial news pages
• Junior Mining Network — preserves "Newsfile Corp." and "ACCESS Newswire" datelines verbatim
• BioSpace — preserves wire datelines for biotech releases

Carriers that bury attribution:

• Yahoo Finance — strips the wire dateline in many cases, presents as "Yahoo Finance"
• CNBC, Variety, Quartz, MacRumors, Futurum — original editorial; wire is not surfaced as source
• Company IR pages without dateline preservation — Apple, Microsoft, AST SpaceMobile all host releases without wire credit

The implication is buyable infrastructure for any wire. A wire that aggressively syndicates into brand-preserving carriers (Morningstar, Stocktitan, FinancialContent, QuiverQuant) buys itself AI visibility independent of the engine's primary retrieval choice. ACCESS Newswire's apparent strength in junior mining is downstream of its Morningstar partnership preserving the "ACCESS Newswire" brand in every URL. Newsfile's strength is downstream of StreetInsider, FinancialContent, and Junior Mining Network all preserving the "Newsfile Corp." brand.

This is also a measurable diagnostic. A Citation Audit can map which carriers preserve attribution for any given wire and identify where the syndication graph leaks. The publication-side view of the same question — which domains carry the most citation weight overall — is ranked in The AI Platform Citation Source Index 2026.

---

Finding 4 — M&A Fragments Wires by Deal Phase

Same deal, different wires across announcement stages. Observed directly:

• Prosperity Bancshares / Stellar Bancorp. January 2026 announcement via PR Newswire. April 2026 regulatory approval release via Business Wire. Same deal. Two wires. Both surface directly inside the engines.
• Talkspace / Universal Health Services. March 2026 deal announcement via PR Newswire. May 2026 shareholder approval via GlobeNewswire.
• Highspot / Seismic. Both sides distributed via Business Wire — but with different syndication patterns depending on which company was the headlining party.

Implication for M&A practitioners. Consistency of wire across the deal lifecycle is a measurable gap. A deal that runs PR Newswire for the announcement and GlobeNewswire for shareholder approval fragments its citation profile inside the engines. The retrieval surface for "What is the status of the [deal name] acquisition?" gets split across both wires' attribution graphs — and the engines weight the most recent release. The first wire's announcement investment gets diluted by the second wire's update.

This is solvable. A locked single-wire strategy across deal phases is a simple Citation Share optimization. Most comms teams do not currently treat wire choice as a deal-lifecycle decision.

---

Finding 5 — Wire Pricing and AI Visibility Are Inversely Correlated

The expensive wires sell to the brands the AI engines won't cite. The cheap wires sell to the brands the AI engines do cite.

Current 2026 wire pricing — verified from primary sources:

Wire	Base 400-word national	Membership	Realistic full cost (1,000 words + image)	Owner
PR Newswire	$850	$195/yr	$1,500–$3,000+	Cision
Business Wire	$760	None	~$2,210+	Berkshire Hathaway
GlobeNewswire	~$350	None	$600–$900 domestic; $1,500+ international	Notified
TMX Newsfile	Quote-based	None	$400–$800 (junior mining specialist tier)	TMX Group
ACCESS Newswire	Starts $415	None	$175–$1,095 range; 5-pack $660 each	Issuer Direct Corp (NYSE: ACCS)

PR Newswire and Business Wire are the most expensive. They also have the highest dead zone rate — used by the brands the AI engines will not surface. GlobeNewswire is the cheapest of the Big Three and the highest in Citation Share. Newsfile and ACCESS Newswire are the cheapest of the five and the only two wires that own a vertical outright.

Cost-per-citation normalization. Applying observed Citation Share to full-cost-per-release across a typical 24-release year illustrates the spread. For a biotech client running 24 releases at $2,210 each on Business Wire (annual cost ~$53,000, observed biotech Citation Share ~0%), the effective cost per direct citation is undefined — the spend produces no measurable AI retrieval surface in the tested vertical. The same client on GlobeNewswire at $750 per release (annual cost ~$18,000, observed biotech Citation Share ~30%) produces a defined cost-per-direct-citation in the low hundreds of dollars. The Business Wire biotech spend is three times higher and produces zero measurable direct retrieval; the GlobeNewswire spend is one-third and wins the vertical. The comparison holds directionally across the other verticals where a wire wins.

The conclusion is sharp. AI visibility is not for sale at the high end of the wire price tier. The premium wires deliver SEC compliance, AP and Reuters reach, and tier-one financial wire credibility — real value, but not AI retrieval value. The cheap wires deliver retrieval surface inside the engines because their client base sits in verticals where editorial does not crowd the wire out.

---

Finding 6 — The Small Wires Are Leading the AI Visibility Arms Race

While PR Newswire and Business Wire continue selling 2010-era value propositions — reach, AP, Reuters, SEC compliance — the smaller wires have publicly repositioned around AI visibility.

TMX Newsfile published primary research on AI crawler activity. Per their analysis: mining and energy press releases generated higher levels of AI activity per release than Technology & Innovation announcements. They studied AI crawler behavior across roughly 120 releases over a 72-hour window. Their study confirms — from the wire's own data — what this audit observed from the retrieval side.

ACCESS Newswire (NYSE American: ACCS) launched ACCESS Insights & Analytics, an AI-powered analytics platform that scores press releases across eight AI metrics and predicts LLM discoverability. The publicly traded parent (Issuer Direct Corp) carries a Strong Buy analyst rating and a 12-month price target of $14 on the bet that AI visibility tooling becomes the wire industry's next product category.

GlobeNewswire (Notified) positions its 2026 pricing pages around AI-ready metadata as a built-in differentiator versus PR Newswire.

PR Newswire (Cision) and Business Wire (Berkshire Hathaway) are quieter on the AI visibility positioning. They continue to lead on AP and Reuters credibility and SEC EDGAR compliance — both real, both valuable, neither solving for AI retrieval.

The small wires are betting their future on the question this study measures. The big wires are betting on the value propositions that built their networks twenty years ago. The audit data suggests the small wires' bet is the right one.

---

Finding 7 — The AI Crawler Access Layer

The wires that let AI engines through the front door win the retrieval surface. The wires that don't, don't.

Observed directly at the robots.txt layer across the five wires' primary release-hosting domains in June 2026:

• GlobeNewswire (globenewswire.com) — permits GPTBot, ClaudeBot, PerplexityBot, Google-Extended, Applebot across the release path. Effectively fully open to AI crawlers.
• TMX Newsfile (newsfilecorp.com) — permits AI crawlers on the release-detail path; TMX Newsfile's own primary research paper explicitly measured AI crawler traffic against their release corpus, which requires open crawl access to conduct.
• ACCESS Newswire (accessnewswire.com) — permits AI crawlers; the ACCESS Insights product roadmap depends on AI-engine discoverability of ACCESS-hosted releases.
• Business Wire (businesswire.com) — mixed signal; certain crawl paths are restricted, and Business Wire has been publicly cautious on AI training-data usage. The observed retrieval surface is downstream of syndication carriers (Morningstar, Yahoo Finance) more than the Business Wire domain itself.
• PR Newswire (prnewswire.com) — historically the most restrictive of the five on crawler access; Cision's institutional posture prioritizes protecting the syndication network's exclusive-content value. This is consistent with the observed pattern of PR Newswire winning direct citation only in the large-cap B2B vertical where downstream carriers preserve attribution.

The pattern maps to the Citation Share leaderboard almost exactly. Wire-domain openness to AI crawlers correlates with wire-domain direct citation. The wires that permit ClaudeBot, GPTBot, PerplexityBot get cited on their own domain. The wires that restrict crawl access get cited only through the syndication carriers that already permit crawlers. This is buyable infrastructure — a wire that opens its release path to AI crawlers can measurably move its Citation Share in the next quarterly re-run of this audit. Phase 2 will quantify the correlation.

Crawl Access is the 5% component of the Citation Share formula for exactly this reason — the full weighting is documented at Citation Share: The New Discoverability KPI.

---

Finding 8 — Wire-Tier Mirrors Client-Tier (The Selection Caveat)

Honesty requires this caveat. Wire dominance correlates with client tier — which means wire choice is partially an artifact of company size and vertical, not pure wire network quality.

Junior miners use Newsfile and ACCESS Newswire because those wires are cheap and dominate the junior mining trade-press syndication graph. Mid-cap biotech uses GlobeNewswire because it is cheap and dominates the biotech editorial graph. Mid-cap medtech and B2B SaaS use Business Wire because the brand carries weight with the syndication carriers that index those verticals. Mega-cap consumer tech uses PR Newswire and Business Wire for SEC and AP and Reuters distribution — and gets eaten by editorial because that is what mega-cap consumer tech always gets.

This does not invalidate the Citation Share findings. Companies experience the wire choice they make — the observed retrieval matters regardless of cause. But a controlled experiment — same release distributed via different wires — would isolate network quality from client-tier effects. That is Phase 3.

For now, the actionable implication is unchanged. Whether GlobeNewswire's biotech dominance is caused by network quality or client-base selection, a biotech client publishing through PR Newswire instead of GlobeNewswire will see lower Citation Share inside Claude and Google AI Overviews — measurably.

---

What Comms Teams Should Do Monday Morning

Audit the dead zone. If your brand is mega-cap consumer tech, calculate annual wire spend, calculate AI engine Citation Share, and report the gap. If the wire spend has SEC and AP and Reuters justifications, document them. If the only justification is AI visibility, the spend is a sunk cost.

Map your syndication carriers. Identify which downstream carriers — Morningstar, Stocktitan, FinancialContent, Yahoo Finance, BioSpace, Junior Mining Network — currently preserve your wire attribution and which strip it. Negotiate, syndicate, or pay into the brand-preserving carriers.

Pick the wire that owns your vertical, not the wire your CFO recognizes. A biotech client paying $2,210 for Business Wire is overpaying versus a $600 GlobeNewswire release that wins more retrieval surface inside the engines. A junior miner using PR Newswire is buying nothing — the entire vertical is owned by Newsfile and ACCESS Newswire for one-fourth the cost.

Lock single-wire strategy across M&A deal phases. Multi-wire distribution across announcement, regulatory approval, shareholder vote, and close fragments the retrieval graph. Single-wire consistency consolidates Citation Share for the "What is the status of the [deal] acquisition?" query.

Ask your wire vendor about their crawler-access policy. A wire that restricts GPTBot, ClaudeBot, and PerplexityBot at the robots.txt layer is structurally handicapping every release you distribute through it. This is a procurement-tier question that most CCOs have not thought to ask; every wire's institutional stance on AI crawler access is now a material variable in AI Citation Share outcomes.

Treat Citation Share as the metric, not pickup count. Wire reports measure media list size and pickup count — both 2010-era metrics. Citation Share inside Claude, ChatGPT, Gemini, Perplexity, and Google AI Overviews is the only metric that maps to how buyers, journalists, and analysts actually research in 2026. Baseline your own number with the 5-step AI Visibility Audit.

Recognize that editorial owns the answer layer. The biggest implication of this study reaches past wire services. AI engines retrieve editorial rewrites, not primary-source releases. That changes the prioritization of every press operation. Earned media coverage from CNBC, Reuters, Bloomberg, BioSpace, or Stocktitan is now structurally more valuable than the underlying release. The release exists to trigger the rewrite. The rewrite is what the engines cite.

---

What Comes Next

Phase 2 expands the audit to 200 prompts across the five verticals, with direct testing across all five engines via API. Five wires. 250-release reference corpus. 1,000 retrieval cells. Per-engine, per-vertical, per-wire Citation Share matrix. Quantitative syndication carrier graph. Decay curve measuring citation rate at day 1, day 7, day 30, and day 90. Ships Q3 2026.

Phase 3 is the controlled distribution experiment — the same release distributed across multiple wires to isolate network quality from client-tier selection. The wire industry has never let anyone run this study. We will try.

The Wire Service Citation Audit will be published annually. Citation Share is the new market share for press release distribution.

---

Methodology Notes and Limitations

This is the first annual edition of the Wire Service Citation Audit. Honesty about scope:

• Sample size. 24 prompts is directional, not definitive. Patterns are consistent across verticals — but a sophisticated reader should read the findings as hypotheses confirmed at pilot scale, with Phase 2 designed to test them at 200+ prompts.
• Two engines tested directly. Claude and Google AI Overviews. ChatGPT, Gemini, and Perplexity are covered in Appendix A via retrieval-architecture analysis only.
• Client-tier selection bias. Wire dominance correlates with client mix. The findings measure observed retrieval, not isolated wire network quality. Phase 3 isolates this.
• Prompt-set bias. Buyer-research style prompts may favor wires that distribute commercially framed releases. Technical, legal, and regulatory prompts may yield different patterns.
• Time bias. This study ran in June 2026. AI engine retrieval models update continuously. The audit will be re-run quarterly going forward.

Reproducibility. Full prompt set, release corpus, scoring rubric, and per-prompt data table are available on request for journalists, researchers, and wire service teams seeking to replicate or challenge the findings. The scoring framework is documented at The GEO Citation Audit: Six-Step Methodology.

---

Appendix A — Retrieval-Architecture Analysis: ChatGPT, Gemini, and Perplexity

This study tested Claude and Google AI Overviews directly. ChatGPT, Gemini, and Perplexity are analyzed structurally using public web index signal — the same source layer those engines retrieve from. This analysis extrapolates from observed indexing patterns; it is not a substitute for direct API testing.

ChatGPT (OpenAI). Retrieves heavily from Bing's search index. Bing surfaces the same syndication carriers observed in the Google data — Morningstar, Stocktitan, BioSpace, FinancialContent — at comparable density. ChatGPT's source citation behavior favors authoritative editorial (CNBC, Reuters, Bloomberg) and primary-source domains (SEC EDGAR, company IR pages) for financial queries. Wire domain direct citation appears for the same vertical-locked patterns observed in Google AIO — small/mid-cap biotech (GlobeNewswire), B2B SaaS (Business Wire), junior mining (Newsfile and ACCESS Newswire). Mega-cap consumer tech surfaces editorial only. Expected ChatGPT Citation Share pattern: substantially identical to the Google AIO findings.

Gemini (Google). Built on the same Google search infrastructure as Google AI Overviews. Differences in retrieval are model-side, not source-set-side. Expected Gemini Citation Share pattern: closely mirrors Google AI Overviews findings, with marginal variance in how the source set is summarized.

Perplexity. Indexes both general web (Bing and proprietary) and specialty sources. Source citation behavior favors primary-source domains more aggressively than ChatGPT or Google AIO — Perplexity is more likely to cite the wire domain directly when it appears in retrieval, less likely to substitute editorial rewrites for primary sources. Expected Perplexity Citation Share pattern: higher direct wire-domain citation rate than the other engines in verticals where wires win (biotech, medtech, junior mining). Mega-cap dead zone effect still applies — Perplexity also cites editorial for Apple, Meta, Microsoft, Nvidia coverage.

Phase 2 closes this gap with direct testing across all three engines via API.

---

The Citation Audit Family

This study is one applied audit in a series. Each runs the same six-step methodology against a different source layer.

• The GEO Citation Audit: Six-Step Methodology — the framework every audit in this family runs against.
• The IR Page Citation Audit 2026 — corporate investor relations pages. Financial services ~100%, biotech ~92%, mega-cap tech and junior mining ~0%.
• The Vertical Citation Audit 2026 — seven verticals: luxury hotels, universities, fashion, hospitals, restaurants, law firms, financial advisors.
• LLM Citation Audits: 100 Consumer Brands — legacy brands underperform digital challengers by 32% on Citation Share.
• The Podcast Citation Index 2026 — which podcasts the five engines actually cite, ranked.
• The Paywall Penalty 2026 — hard-paywall publishers captured 0% of AI-retrieval citations.

Brand-side companion: AI Visibility Audits · The AI Visibility Audit: 5 Steps · Citation Share: The New Discoverability KPI

The complete research catalog: Everything-PR Research: The Master Index

---

About This Study

The Wire Service Citation Audit 2026 is independent research published by Everything-PR. Methodology, prompt set, release corpus, and per-prompt data tables are available on request for replication, challenge, or peer review.

---

Related: Business Wire: The Berkshire Hathaway-Owned Newswire · The PR Newswire Sale: How Cision Bought the Wire · The Wire Services in Asia · Fake Press Releases: How Wire Service Hoaxes Manipulate Markets · Cision and PR Newswire