I Would Not Charge International Press to AI Visibility

I would not pay for international press out of the AI visibility budget. I would pay for it out of market entry, and I would judge it on pipeline in that market. The pitch that a London or Berlin placement is also how an answer engine learns the company exists there does not survive the country cut of the citation index. Coverage can still be worth buying for the people who read it. It does not get a second job it has not earned.

The dataset is Country-Code Domains Take 2.7% of AI Citations | MR Research. That page is the deeper resource. This essay is how I would use it as a founder: which budget line the placement sits on, which outlets I would still buy for human readers, and which release would move the money back.

What is true about the human reader

International coverage still does the job it always did for a person in that market. The Reuters Institute's 2026 Digital News Report, across 48 markets, finds trust in news at the lowest level it has recorded since 2015, and finds trust in the most widely used individual news brands holding up better than trust in news as a whole (Reuters Institute). A reader in Sydney who reads a specific paper still weights that paper. Cision's 2026 State of the Media, 1,899 journalists across 19 markets, finds 66% of them relying on PR-provided material for story ideas (Cision). The placement is winnable, and the reader is real. I would not argue with either fact.

The second half of the pitch is the part I measured. The claim is that the same placement is how an AI answer reaches that market. That is a claim about which domains the engines cite, and it can be checked.

What the release says about where domains are registered

Machine Relations Research published the country-of-registration cut from release mri_score_v2.0+2026-09-19+0cad03121f60, window 2026-05-10 to 2026-09-19, 15,883 answer runs, 22,213 cited domains, 111,510 domain-run citation pairs, six engines. Every figure in this section is from that release and from the country-code page that reads it.

Where the cited domain is registered	Domains	Share of citation volume
Generic (.com, .org, .net and kin)	17,876	85.17%
Two-letter codes sold as brandable strings (.ai, .io, .co)	2,611	10.45%
National country codes, 91 territories	1,462	2.73%
US-restricted (.gov, .edu, .mil)	264	1.65%

The Index publishes a segment rate only after 10 observations across 7 distinct run dates, and grades a domain only once enough evidence stands behind it. Three of the 1,462 national domains carry a grade, and all three are American businesses using .us as a short name. Non-US national domains with a grade: zero, out of 1,410.

The whole .uk zone, 337 cited domains, is 0.600% of citation volume. The whole .de zone, 96 domains, is 0.133%: 148 citation pairs out of 111,510 across 126 days, from a registry DENIC reports as one of the largest in the world (DENIC). Japan contributed 22 domains and 34 pairs. Meanwhile .ai, delegated to the Government of Anguilla (IANA), and .io, delegated in the British Indian Ocean Territory (IANA), take 8.56% between them, which is 3.13 times every nation combined. Both are country codes under RFC 1591 and ICANN's definition. Neither says where a company operates.

I would not hear "country code" and think "local press." A two-letter suffix is a registry fact. On this release the brandable two-letter suffixes carry more of the answer layer than the national press suffixes do. That is the source layer. The outlets are the next check, because a founder is not pitched a suffix. A founder is pitched a newsroom.

The 45 outlets, read one by one

I put 38 non-US outlets a founder is routinely pitched, across the UK, Europe, Asia-Pacific, Canada, and India, next to 7 US outlets for scale, and I read each public profile on 2026-09-19 at 13:52 UTC. The route is machinerelations.ai/index/domains/ plus the domain. A profile lists cited runs, days cited, engines, and a confidence grade. No profile means the domain is not among the 22,213 cited at least once in the window.

These are the non-US outlets that had a profile, ordered by cited runs out of 15,883.

Outlet	Cited runs	Days cited (of 126)	Engines (of 6)	Grade
indiatimes.com	37	24	5	C
theguardian.com	29	10	3	Collecting
thenextweb.com	20	14	5	Collecting
reuters.com	13	13	3	Collecting
independent.co.uk	12	8	2	Collecting
smh.com.au	10	6	1	Collecting
sifted.eu	7	7	4	Collecting
theglobeandmail.com	6	5	3	Collecting
bbc.co.uk	5	3	1	Collecting
tech.eu	5	5	2	Collecting
bbc.com	4	3	1	Collecting
ft.com	4	4	2	Collecting
lemonde.fr	4	4	1	Collecting
afr.com	4	4	1	Collecting
economist.com	3	3	2	Collecting
cbc.ca	3	2	2	Collecting
handelsblatt.com	2	2	1	Collecting
elpais.com	2	2	1	Collecting
thetimes.com	1	1	1	Collecting
eu-startups.com	1	1	1	Collecting
techinasia.com	1	1	1	Collecting

Twenty-one outlets, 173 cited runs between them, one grade. The other 17 returned no profile that day: aljazeera.com, corriere.it, dw.com, euronews.com, faz.net, france24.com, lesechos.fr, nikkei.com, scmp.com, spiegel.de, straitstimes.com, telegraph.co.uk, theaustralian.com.au, thehindu.com, thetimes.co.uk, wired.co.uk, and zeit.de. Der Spiegel, Nikkei, the South China Morning Post, and the Telegraph are not small newsrooms. Across 15,883 English-language commercial answers, the six engines did not cite any of them once in this window.

A grade is domain-level. The Index assigns one only after a domain has enough evidence across its segments. In this release, 503 of 22,213 cited domains carry a grade, and no graded domain has fewer than 30 cited runs. One non-US newsroom on the list has a grade: indiatimes.com, at C, cited by five of the six engines. The Guardian has 29 runs on three engines and still sits at Collecting, because those runs are spread thinly across six segments in three categories.

The seven US outlets, same day, same release:

Outlet	Cited runs	Days cited	Engines	Grade
forbes.com	657	105	6	A
techcrunch.com	164	44	5	B
nytimes.com	141	31	3	B
cnbc.com	129	31	5	B
businessinsider.com	99	40	5	C
wsj.com	74	23	4	C
bloomberg.com	3	3	2	Collecting

Forbes alone is cited in 657 runs. The 21 non-US outlets with a profile are cited in 173 between them. One American business magazine outweighs that combined non-US set by 3.8 to 1 in this answer layer. Bloomberg, at 3 runs, is the reminder that prestige is not the mechanism either. The mechanism is which pages the engines retrieve on a commercial question. I would not buy the Telegraph for citations because the Telegraph is important. On this release it was not cited once.

News questions are the least international

If any question shape reached into national press, I would have expected the news-shaped one. The Index measures seven shapes. National-domain share is lowest on news.

Question shape	Citation pairs	National-domain share
Problem-first research	8,856	4.55%
How buyers choose	11,242	3.93%
Ranked top lists	12,774	3.59%
Best tools	12,713	3.27%
Head-to-head comparison	9,915	2.97%
Is it worth it	10,794	2.81%
News-driven	45,216	1.61%

A problem-shaped buying question is 2.8 times more likely to reach a national domain than a news question about the same industry. Christian Lehman, Chief Growth Officer, published the companion finding the same day: the publications an engine cites for an industry's news are not the ones it cites for its buyers. The international pitch assumes news is the globally distributed surface and product content is the local one. On these shares it runs the other way. News-driven questions are 45,216 of the pairs and only 1.61% national.

Category moves the national share more than the country on the pitch

Category	National-domain share	.ai and .io share
Emergent prosumer	7.15%	4.30%
Consumer health	6.91%	1.68%
Consumer products	6.27%	1.23%
Cybersecurity	1.93%	10.32%
Enterprise software	1.80%	8.75%
Fintech	1.64%	10.63%
AI visibility and GEO	2.44%	26.17%

Where the answer depends on a regulator, a health service, or a physical product sold in a place, the engines reach national infrastructure: nhs.uk, europa.eu, cdhf.ca, vic.gov.au. The most-cited non-US national domain in the release is europa.eu, cited by all six engines, and it is a government, not a publisher. Where the answer is B2B software, national domains barely appear and .ai and .io vendor sites take the space. In AI visibility and GEO, brandable two-letter domains take 26.17% of citation volume against a national share of 2.44%.

A consumer-health founder and a fintech founder hearing the same international pitch are being sold two different goods. For one, the national layer is already in the answer, and it is a public institution. For the other, the national layer is a rounding error next to vendor domains. I would open the category segment, count national domains in the top ten, and only then decide whether a national paper is aimed at the answer or at the reader.

The order I would build either way is Machine Relations. I keep the registration cut and the grade rules from sliding into ordinary PR vocabulary by using the glossary for the terms, so "national" does not quietly come to mean "a newspaper in that country."

How I would answer the pitch in the room

The pitch arrives as a list of newsrooms and a sentence about algorithms learning the market. I would answer with the release id on the table, mri_score_v2.0+2026-09-19+0cad03121f60, and I would read the pitched domains against the census I already took on 2026-09-19 at 13:52 UTC. If the domain had no profile, it was not among the 22,213 cited domains in 15,883 runs. If it had a profile and no grade, it has not accumulated the evidence the Index requires for a grade. Only 503 of those 22,213 domains carry one, and each of those has at least 30 cited runs. indiatimes.com is the exception on my list of 38, grade C, 37 runs, 24 days, five engines. The Guardian at 29 runs on 10 days and three engines is still Collecting. I would say that out loud, because "the Guardian" is the name that makes the rest of the list feel safe. On this window it is a thin footprint, not a strategy.

I would then show the registration cut so the conversation cannot hide inside a famous masthead. Generic domains, 17,876 of them, carry 85.17% of citation volume. Brandable two-letter suffixes, 2,611 domains, carry 10.45%. National codes across 91 territories, 1,462 domains, carry 2.73%. US-restricted suffixes carry 1.65% from 264 domains. Three national domains carry a grade, and all three are .us used as a short name. The non-US remainder with a grade is zero of 1,410. The .uk zone's 337 domains at 0.600%, and the .de zone's 96 domains at 0.133%, 148 pairs out of 111,510, are the local-press theory stated as a share. .ai and .io at 8.56% together, 3.13 times every national zone combined, are why a startup on a brandable suffix is not "international PR." The suffix is a registry fact. Anguilla and the British Indian Ocean Territory are the delegations. They are not markets we are entering.

The shape table is how I would stop a news-placement from being described as the globally distributed surface. News-driven questions are the largest block, 45,216 citation pairs, and the least national, 1.61%. Problem-first research is 4.55% national on 8,856 pairs, 2.8 times the news share. A firm that sells international news coverage as AI distribution is selling the shape where national domains are weakest. Christian Lehman's finding from the same day, that news citations and buyer citations are different publication sets, is the sentence I would borrow. I would not borrow it as a reason to skip a newspaper a human reader trusts. Reuters, across 48 markets, still finds people weighting the brands they actually use, even while trust in news as a category sits at a recorded low since 2015. Cision's 1,899 journalists, 66% of them using PR material for ideas, is why the placement can be won. I would win it on the market-entry line.

Category is the last page of the answer. Consumer health at 6.91% national and fintech at 1.64% national, with .ai and .io at 10.63% in fintech and 26.17% in AI visibility, means I would not let two portfolio companies share an international citation budget. Where the answer needs a government, europa.eu is the most-cited non-US national domain and it is cited by all six engines. A newspaper is not a substitute for a regulator the engines are already reading. Where the answer is software, I would open the category segment and count national domains in the top ten before I approved the flight and the clipping.

On the strict reading that calls every two-letter suffix national, .ai, .io, and .co are still 75.5% of that group. I would mention the strict reading so a critic cannot accuse the split of being a labeling trick. The 17 outlets with no profile, including Der Spiegel, Nikkei, the South China Morning Post, and the Telegraph, keep the trick from being the point. They were not cited once in this English commercial basket. Forbes, at 657 runs over 105 days on six engines, grade A, outweighs the 173 runs of the 21 profiled non-US outlets by 3.8 to 1. Bloomberg at 3 runs is there so I do not sneak prestige back in through a US door. I would fund the reader. I would not fund the suffix.

What I would do

I would take international coverage off the AI visibility line and leave it on market entry. The humans in that market still weight the brands they use, which is the Reuters finding, and the journalists are reachable, which is the Cision finding. The success condition is pipeline in that market. Zero graded non-US national domains, and one graded newsroom out of 38, is not a citation line. It was two line items sharing a sentence.

For citation outside the United States I would spend where this release already shows the engines reaching, which for software categories is the generic layer: review sites, community sites, aggregators, and the .com, .ai, and .io domains sitting at the top of the category segment. I would open that segment before I approved a national paper as a citation tactic. I would not move the site onto a country-code domain in order to be cited. Google's documentation treats a ccTLD as a strong signal of country targeting (Google). That is the property that hides a domain from an engine answering a question with no country in it. The .de zone's 148 pairs out of 111,510 are the scale of that hide. A company that needs local pages can keep localized versions on the domain the global layer already cites (Google).

I would not fund a claim this instrument cannot see. The release carries no locale, language, or region field. It observes English-language commercial questions with no country in them. It cannot tell me what a German buyer asking in German sees. Unmeasured is not the same as zero, and unmeasured does not get the visibility budget. It gets a measurement first.

On the strict reading, where every two-letter domain counts as national, .ai, .io, and .co are still 75.5% of that group. The editorial split between brandable suffixes and national suffixes is reported both ways. I would not let a debate about the label rescue a newspaper the engines did not cite.

A further descriptive fact from the same release, for anyone tempted to "just get one national domain in": 86.5% of national domains were cited by exactly one engine, and 61.6% in exactly one run. A single appearance is not a strategy. Forbes at 657 runs across 105 days and six engines is what a repeated selection looks like in this window. The Guardian at 29 runs on 10 days is a real newsroom having a thin answer-layer footprint, which is a different fact from the newsroom being unimportant to its readers.

Scope: English commercial questions with no country in the prompt, six engines, 126 days, release mri_score_v2.0+2026-09-19+0cad03121f60. Citation here is an observation of what was retrieved. I would not publish that a placement caused or failed to cause a citation. The machine-readable release is at machinerelations.ai/data/machine-relations-index.json. Classification is by the last label of the domain, which is what a registry and a resolver both use.

What would change my mind: any later release in which 10 or more non-US national domains carry a grade, where today the count is 0 of 1,410; or any outlet on the 38, other than indiatimes.com, carrying a grade of B or better; or a locale-varying measurement that shows a national-domain share above 10% for a market I actually sell into. If one of those fires, I would move a defined slice of citation budget toward that market's domains and I would write the new release id next to the old one. Until then the review is the same 45 profiles, read again on the next release, at the same routes.

FAQ

Does this mean I should stop buying international PR?

No. I would buy it for readers in that market and judge it the way market entry is judged, on pipeline there. The Reuters report and the Cision survey are the case for the human job. The case they do not make is a second charge to AI visibility. Zero of 1,410 non-US national domains carry a grade on this release.

Would the Guardian or the FT change my AI citations?

On this window the Guardian was cited in 29 runs and the FT in 4, and neither has a publishable grade. Forbes was cited in 657. I would pursue either paper for its readers. I would not write the purchase order as a citation strategy unless a later release moves the grade.

My buyers are in Germany. Does this say what they see?

It says what an English question with no country in it retrieves. The .de zone's 148 citation pairs describe that basket. They do not describe a German-language query asked in Germany. I would run that measurement before I spent citation money on a German newsroom. I would not pretend this release already ran it.

Should I move the site to a country-code domain for the new market?

Not for citation. Google treats a ccTLD as a strong country signal, and this release cites national domains rarely: 86.5% by exactly one engine, 61.6% in exactly one run. I would keep localized pages on the domain the engines already cite, and I would let the country-code decision be a market-entry decision with a market-entry goal.

How would I check the next outlet a firm pitches?

I would open the domain profile for that root domain and read cited runs, days, engines, grade, and the release id at the top. A missing profile means the domain was not cited in that release's 15,883 runs. I would read it the day I decide, because a later release is allowed to change the count, and the change is the next decision, not a reason to ignore the current one.