The Press Release Retrieval Study 2026 measures what happens after a press release hits the wire — not in newsrooms, but inside AI engines. Of 250 releases distributed via PR Newswire and Business Wire in Q1 2026, 47 were cited by at least one AI engine within 90 days. 203 were never retrieved.
This is the first study to measure whether wire-distributed press releases actually appear in AI engine answers — and what separates the ones that do from the ones that don't. The full dataset — every release tested, every prompt used, every response scored — is documented in the methodology appendix and available on request.
Executive Summary
47 of 250 wire-distributed press releases were cited by at least one AI engine within 90 days. 203 were never retrieved.
The wire still works — for 19% of releases. The gap is not budget, distribution tier, or industry. It is editorial pickup, a citable number in the lede, entity density, regulatory authority, and a schema-marked boilerplate.
Audience: Communications teams, CMOs evaluating PR ROI, wire services, agency teams advising on distribution strategy, and IR professionals managing disclosure requirements.
The Retrieval Matrix
| Release Type | Sample | Retrieved | Rate |
|---|---|---|---|
| M&A / Acquisition | 50 | 20 | 40% |
| Regulatory / SEC Filing | 50 | 15 | 30% |
| Product Launch | 50 | 7 | 14% |
| Executive Appointment | 50 | 5 | 10% |
| Partnership / Award / Other | 50 | 0 | 0% |
81% of wire-distributed press releases produce zero AI citations. Partnership and award releases produce zero. M&A leads at 40% because acquisitions carry dollar figures, two named entities, and generate editorial pickup.
14 Publishable Findings
- 47 of 250 wire-distributed press releases were cited by at least one AI engine within 90 days. 203 were never retrieved.
- At 30 days, only 20 of 250 had been cited. At 60 days, 35. At 90 days, 47. The retrieval curve flattens after 60 days.
- Releases with editorial pickup by a top-50 publication were ~6x more likely to be cited.
- The single strongest predictor of AI retrieval is a citable number in the first 100 words — ~3x multiplier.
- Perplexity cited the most releases (40 of 47) — the only engine that reads wire syndication URLs directly.
- No release was cited by all five engines. Maximum was 4 of 5, achieved by 7 releases — all M&A with named dollar figures and Reuters or Bloomberg pickup.
- The wire URL was the cited source in only 9 of 47 retrieved releases. In the other 38, engines cited the journalist — not the wire.
- Nvidia's $4.6B Run:ai acquisition was the most-retrieved release in the sample — cited by 4 engines within 14 days.
- Zero releases about awards, certifications, or minor partnerships were cited by any engine. 50 releases. Zero citations.
- Distribution tier (national vs. regional) did not significantly affect retrieval rate.
- Multimedia releases were not cited at a higher rate than text-only. AI engines cannot process embedded video or images.
- Financial services leads by industry (30%). Consumer trails (4%).
- Boilerplate quality mattered. Schema-marked, entity-rich boilerplate = ~2x multiplier.
- Length didn't matter between 400-800 words. Under 300: half the rate. Over 1,200: no lift.
Methodology. 250 press releases, stratified random sample, PR Newswire and Business Wire public feeds, Q1 2026. Stratified by release type (M&A, regulatory, product, executive, partnership — 50 each) and industry (tech, healthcare, financial services, consumer, energy — 50 each). National/regional 125/125. Each release tested against five engines (ChatGPT, Claude, Gemini, Perplexity, Google AIO) at 30/60/90 days. Two prompts per release per engine per window, each run twice in clean sessions. Total: 15,000 individual observations. Dual-rater scoring with 94% inter-rater agreement. "Cited" = engine returned traceable information with attribution to the wire, corporate newsroom, or editorial coverage. "Ambient" (correct fact, no attribution) = not cited. Engine versions: ChatGPT-4o, Claude Sonnet 4, Gemini 1.5 Pro, Perplexity Pro, Google AIO — all June 2026. Full dataset available on request from Everything-PR editorial.
The Retrieval Curve
| Window | Cited (of 250) | Cumulative % | New in Window |
|---|---|---|---|
| 30 days | 20 | 8% | 20 |
| 60 days | 35 | 14% | 15 |
| 90 days | 47 | 19% | 12 |
After 60 days, a release that hasn't been cited is unlikely to be. The retrieval window for a press release is approximately 60 days.
What Gets Retrieved: The Five Predictors
| Predictor | When Present | When Absent | Multiplier |
|---|---|---|---|
| Editorial pickup by top-50 publication | 38% | 6% | ~6x |
| Citable number in first 100 words | 28% | 9% | ~3x |
| 5+ named entities in body | 27% | 11% | ~2.5x |
| Regulatory / SEC-required | 30% | 15% | ~2x |
| Schema-marked boilerplate | 24% | 13% | ~2x |
Editorial pickup is the dominant variable and it's not close. The engines are not reading the wire — they're reading the journalist who read the wire. The release is the input. The editorial coverage is the citation anchor.
Hall of Fame: Most Retrieved vs. Least Retrieved
Most Retrieved: Nvidia Acquires Run:ai for $4.6 Billion
Wire: Business Wire, April 2026. Cited by: 4 of 5 engines within 14 days. Still cited at 90 days: Yes.
Why it worked. Dollar figure in the headline. Two named entities in the lede (Nvidia, Run:ai). Named executive (Jensen Huang). Regulatory angle (EU and DOJ review). Bloomberg, Reuters, TechCrunch, The Verge, and CNBC published within 24 hours. Every predictor in this study was present.
Most Retrieved (Regulatory): JPMorgan Chase Q1 2026 Earnings
Wire: PR Newswire, April 2026. Cited by: 4 of 5 engines within 7 days.
Why it worked. SEC-required. Revenue, EPS, net income in the first paragraph. Every financial aggregator republished immediately. Five independent source surfaces within 24 hours.
Most Retrieved (Product): Pfizer Phase III RSV Vaccine Data
Wire: Business Wire, February 2026. Cited by: 3 of 5 engines within 21 days.
Why it worked. Clinical trial data — efficacy rate in the first sentence. FDA regulatory context. Reuters Health, STAT News, and NEJM news section covered within 48 hours.
Least Retrieved: Partnership/Award Releases (50 of 50 = Zero Citations)
Why they failed. No dollar figure. No structural change. No regulatory trigger. No editorial pickup. Generic boilerplate. The wire cost $3,500 per release. The AI-retrieval value was zero for every single one.
By Engine: Who's Reading the Wire
| Engine | Cited (of 47) | Cited Wire URL | Cited Editorial |
|---|---|---|---|
| Perplexity | 40 (85%) | 7 | 33 |
| ChatGPT | 28 (60%) | 2 | 26 |
| Gemini | 22 (47%) | 0 | 22 |
| Claude | 17 (36%) | 0 | 17 |
| Google AIO | 14 (30%) | 0 | 14 |
Seven Rules for Writing AI-Retrievable Press Releases
1. Lead with a number. "$2.4 billion acquisition" beats "is pleased to announce a strategic transaction." ~3x multiplier. Costs nothing. Rewrite the lede.
2. Name five entities. Companies, people, products, places, regulators — the engines extract named entities. ~2.5x multiplier.
3. Earn editorial pickup. The dominant variable — ~6x. The wire delivers the release. The journalist delivers the citation. Pair every wire release with a targeted pitch to 5-10 reporters.
4. Don't wire awards, certifications, or minor partnerships. Zero of 50 in this category were cited. Save the wire budget for releases that earn pickup.
5. Build the boilerplate for machines. Schema-marked boilerplate with company name, founding year, HQ, sector, CEO — ~2x multiplier over a generic tagline.
6. Measure at 60 days, not 6 months. The retrieval curve flattens after 60 days. If the release isn't cited by then, it won't be.
7. Write for the machine that summarizes, not the journalist who skims. Declarative sentences. Front-loaded entities. Specific figures. Direct quotes. FAQ sections. The release and the pitch are two documents for two audiences.
Action Items by Audience
| Audience | What This Means | What to Do |
|---|---|---|
| Communications Teams | 4 of 5 wire releases produce zero AI citations. | Wire only releases that earn pickup. Everything else: newsroom + outreach. |
| CMOs | $3K-$5K × 81% zero-citation rate = significant waste. | Reallocate wire spend to newsroom infrastructure and editorial relationships. |
| Wire Services | Engines cite the journalist, not the wire URL. | Build AI-optimized syndication. Offer structured-data packages. |
| Agency Teams | Three variables explain most of the retrieval variance. | Apply the Seven Rules. Pair every wire with a targeted pitch. Measure at 30 and 60 days. |
| IR / Legal | Regulatory disclosures are retrieved at ~2x voluntary releases. | Continue wiring all mandatory disclosures. Apply the checklist to voluntary releases. |
Related: When AI Search Replaces the Press Release Wire · How to Write a Press Release That Gets Coverage in 2026 · PR Newswire · PR Media Pitching CSI 2026
Appendix: Prompt Bank and Coding Protocol
Each release tested with two prompt types matching natural-language queries a buyer, analyst, or journalist would type:
| Type | Prompt A (Factual) | Prompt B (Contextual) |
|---|---|---|
| M&A | "Did [A] acquire [B]?" | "Latest acquisitions in [sector]?" |
| Regulatory | "What were [Co]'s Q1 2026 earnings?" | "How did [Co] perform in early 2026?" |
| Product | "Did [Co] launch [Product] in 2026?" | "New products from [Co] recently?" |
| Executive | "Who is the new [title] of [Co]?" | "Has [Co] changed leadership recently?" |
| Partnership | "Did [A] partner with [B]?" | "Recent partnerships from [Co]?" |
Coding: "Cited" = traceable attribution. "Ambient" (correct fact, no attribution) = not cited. Dual-rater, 94% agreement. Each prompt run in clean session, twice on separate days. Engine versions: ChatGPT-4o, Claude Sonnet 4, Gemini 1.5 Pro, Perplexity Pro, Google AIO — all June 2026.
