What this sample shows
This is a redacted public sample of a single Vertical Intelligence (Complete) deliverable — a snapshot read of the pay-on-results / guaranteed-media-placement PR vertical, prepared in June 2026. The buyer is a UK-anchored pay-on-results PR agency, referred to throughout as the Subject; the Subject's name and its two domains are redacted. The competitive, organic, and demand findings — which describe a public market rather than the buyer — are preserved in full.
What a prospective buyer needs in order to evaluate the product is intact: the three-instrument method (grounded AI-citation probes, an organic-SERP read by query intent, and a tested search-demand signal), the per-finding confidence discipline, and the scoped diligence boundaries that mark exactly what this snapshot does and does not settle.
The verdict
The Subject is in the conversation — at the margins of an answer set owned by SEO-engineered vendors.
When buyers ask an AI engine how to get press coverage, the answer set is not led by PR agencies doing the work — it is led by a sprawl of search-optimised PR-vendor sites and Reddit threads. The Subject does surface: Gemini cites it, in the long tail. But Claude does not cite it at all, and across both engines the most-cited players are vendor sites engineered to be cited — led by authoritytech.io, which appears in both. On the organic side the same pattern holds: Reddit, paid placements, AI Overviews, and job boards intermediate the category's best queries. The Subject is present across these surfaces, but nowhere near the front of them.
- The AI answer set is owned by citation-engineered vendors, not agencies. On the two engines we could ground, the leaders are placepaymedia.com (Claude, 10 citations) and a Reddit-plus-vendor tail (Gemini); authoritytech.io is the only site cited across both engines — and it ranks itself #1 while being a pay-for-performance agency. The answer set rewards sites built to be cited.
- The Subject has a toehold, not a position. Gemini cites both of the Subject's domains — but once each, deep in a 109-citation tail (across 89 domains). Claude does not cite the Subject at all. So the Subject is present in the AI conversation, but never as a leading answer, and absent from one of the two grounded engines entirely.
- The organic surface is intermediated, and demand is soft. Reddit owns agency-recommendation queries; 'SEO services' pushes organic to rank 9+ behind paid and a local pack; head terms now carry AI Overviews. Of the head terms with a significant 12-month signal, 'digital media placement' is genuinely declining (p=0.037, small effect); 'prowly pr' is statistically significant but flat-sloped and brand-specific (the Prowly product, not category demand), so it is not reliable category evidence.
In one line: the Subject is cited, but the category's answer set is won by sites engineered to be cited — and the Subject isn't one of them yet.
How to read this report. HIGH means a well-sampled, directly observed signal; MEDIUM rests on a partial instrument or a single capture. The AI-citation findings here are HIGH for the two engines we could ground (Claude and Gemini both performed live web search and returned cited URLs — 32 citations across 13 domains and 109 across 89 domains respectively). ChatGPT is reported as a MEDIUM control: it answered from training data without live search, so it casts no citation vote either way. Where a signal was thin we name it as a diligence boundary rather than guess.
The AI answer set: who the engines cite
Confidence: High (for the two grounded engines).
We asked ChatGPT, Claude, and Gemini the questions a buyer asks — how to get press coverage, how to get featured, who does pay-on-results PR — and recorded which sites each engine cited. Two of the three grounded successfully and performed live web search. Gemini returned 109 citations across 89 domains; its answer set is led by Reddit, journolink.com, and famehero.com, with a long tail of PR-vendor sites (authoritytech.io, prowly.com, imcwire.com). Claude returned 32 citations across 13 domains, led decisively by placepaymedia.com (cited 10 times) and authoritytech.io (4). The single most telling pattern: authoritytech.io is the only site cited across both engines — and it ranks itself #1 while being a pay-for-performance agency. The answer set is not a meritocracy of PR outcomes; it is a leaderboard of citation engineering.
| Engine | Grounded? | Total citations | Distinct domains | Most-cited domains |
|---|---|---|---|---|
| Claude | Yes — live web search | 32 | 13 | placepaymedia.com (10), authoritytech.io (4) |
| Gemini | Yes — live web search | 109 | 89 | Reddit, journolink.com, famehero.com (long tail incl. authoritytech.io, prowly.com, imcwire.com) |
| ChatGPT | No — answered from training data | — | — | Disclosed control; casts no citation vote |
Figure 1 — Who the AI engines cite: grounded answer set by engine and citation count. authoritytech.io is the only domain cited across both grounded engines; the Subject appears in Gemini's tail, not at all in Claude's.
Where the subject sits in the answer set
Confidence: High.
The Subject is not invisible to the AI engines — but it is marginal. Gemini cited both of the Subject's domains once each, deep in a 109-citation tail (89 domains). Claude did not cite the Subject at all. So across the two engines that actually searched the web, the Subject earned two long-tail mentions out of 141 total citations, and zero leading positions. The strategic read for a visibility business is precise: the problem is not that the Subject is unknown to the engines — it is that the Subject is not built, as the leaders are, to be the answer. Closing that gap is a citation-engineering problem, not a brand-awareness one.
| Engine | Subject cited? | Mentions | Context |
|---|---|---|---|
| Gemini | Yes | 2 (both Subject domains, once each) | Deep in a 109-citation tail across 89 domains |
| Claude | No | 0 | Not cited at all |
| Combined | — | 2 of 141 total citations | Zero leading positions |
Figure 2 — Where the Subject sits: long-tail presence in one grounded engine, absence in the other.
Who owns the organic surface
Confidence: High.
The organic picture mirrors the AI one: the category's most valuable queries are intermediated by platforms, not won by agencies. For agency-recommendation queries, Reddit dominates the top organic positions. For 'SEO services,' paid ads plus a local pack push the first organic result to rank 9 or below. Head terms like 'earned media strategy' now render an AI Overview that shrinks the organic click pool. And 'communications consultant' is intent-mixed — LinkedIn and Indeed job listings, not consulting pages. One framing note: 'SEO services,' 'communications consultant,' and the agency-recommendation queries are adjacent marketing queries rather than core pay-on-results-PR terms — this organic-surface read deliberately uses the broader agency orbit the Subject competes in, not only the narrow category head terms. The front of that surface is held by platforms and job boards; the Subject's own organic ranking position could not be established here — the SERP-history probe of the Subject's primary domain returned no ranking data, which we record as a diligence boundary rather than infer a position from.
| Query / intent | Who holds the front of the organic surface |
|---|---|
| Agency-recommendation queries | Reddit dominates the top organic positions |
| 'SEO services' | Paid ads plus a local pack push the first organic result to rank 9+ |
| Head terms (e.g. 'earned media strategy') | An AI Overview renders, shrinking the organic click pool |
| 'communications consultant' | Intent-mixed — LinkedIn and Indeed job listings, not consulting pages |
Figure 3 — Who owns the organic surface, by query intent. The category's best queries are intermediated by Reddit, paid, AI Overviews, and job boards — not by agencies.
Demand and content supply
Confidence: Medium.
Category demand is modest, with only one clean directional signal. Of the head terms with a statistically significant 12-month result, 'digital media placement' shows a genuine declining slope (Mann-Kendall p=0.037, ~−1.7 searches/month, though a small effect). 'Prowly pr' is also significant (p=0.027) but its slope is essentially flat and it is a brand term (the Prowly product), so it must not be read as category-demand decline. 'Get featured in Forbes' shows a positive but non-significant slope.
| Head term | 12-month trend | Significance | How to read it |
|---|---|---|---|
| digital media placement | Declining (~−1.7 searches/mo) | Mann-Kendall p=0.037 (small effect) | A genuine, small category decline |
| prowly pr | Flat slope | p=0.027 (significant) | A brand term (the Prowly product) — not category demand |
| get featured in Forbes | Positive slope | Not significant | Inconclusive |
Figure 4 — Demand signal across head terms. Demand is not collapsing, but it is not a rising tide either; only one term carries a clean directional read.
The content competing for these terms looks abundant but is largely a measurement artefact. 'Get featured in Forbes' returns ~210,000 pages but only ~1,816 from Forbes itself. And the inflated page-counts for 'performance PR' and 'digital media placement' are driven by Indian internship and job-placement sites (e.g. internshala.com, qxpsols.com) where "placement" means a job placement, not a PR / media placement — a semantic mismatch, not real PR-content competition. This is precisely why off-the-shelf SEO page-counts mislead: they count string matches, not competitors. It is also why manual verification earns its keep — the raw page-count would have over-stated the competitive density several-fold. Demand is not collapsing, but it is not a rising tide, and the apparent SERP clutter dissolves once the off-topic placements are filtered out.
What this sample covers, and what a deeper engagement adds
This is the Complete tier — grounded AI-citation across the engines that returned live search, plus organic SERP and demand. A deeper engagement extends past these named boundaries:
- ChatGPT grounding. ChatGPT answered from training data without performing a live web search this pass, so it casts no citation vote. A deeper engagement completes the grounded multi-engine view — Claude and Gemini ground cleanly, and a search-native engine (e.g. Perplexity) can be added; ChatGPT via this source does not perform live search and stays a disclosed control.
- Home-market demand resolution. The demand read is 69 US-only seed-variant keywords; the Subject is UK-anchored, and UK-resolution volumes plus full keyword expansion were not captured.
- Content-gap quantification. Page-count comparisons rest on single low-confidence samples and off-topic-placement contamination; a deeper pass parses and scores the genuinely competing content.
- Citation-path teardown. We identify who the engines cite; a deeper engagement reverse-engineers what the recurring winners (authoritytech.io, placepaymedia.com) share — their backlink, on-page, and structured-data patterns — and turns it into a prioritized plan to close the Subject's gaps. (Backlink and on-page analysis are directly measurable; structured-data is derived from on-page captures.)
To close the boundaries named above — a re-probe to complete the grounded multi-engine AI-citation view, full home-market plus US category demand sizing, and a citation-path teardown of the leading vendors' backlink, on-page, and structured-data patterns with a prioritized plan — reach the ForIntel desk directly at forintel@foragentis.com, or scope a report on the ForIntel order page.
This is a redacted public sample of a ForIntel Vertical Intelligence (Complete) deliverable, published by Foragentis to demonstrate the method. The buyer's identity and domains are redacted; the competitive, organic, and demand findings reflect a public market and are preserved. Findings reflect grounded AI-citation probes, organic-SERP, and search-demand signals captured in a snapshot window; confidence levels and scoped boundaries are stated per finding by design. A Foragentis product · foragentis.com/forintel.