Which agencies AI engines name, measured across four engines

Broadcastwell publishes this page and provides AI search visibility services. The dated measurement table is a research snapshot, not a permanent ranking; current commercial claims and client evidence are presented separately.

answers scored across four engines

of those answers came from questions that named no agency

total namings across those 120 answers

of 223 cited domains appeared exactly once

The shortlist, name-free

Between 22:30 on 26 July and 01:15 on 27 July 2026 we put ten buyer questions about AI search visibility agencies to Claude, ChatGPT, Perplexity and Google AI Overviews, five times over. Forty answers per run, 200 in total. Six of the ten questions named no agency, which leaves 120 answers where the engine had to pick a name unprompted.

Those 120 answers are the ones worth reading. The other 80 came from questions shaped as alternatives-to or one firm versus another, which hand the engine a name and then reward it for repeating it.

Data table: Agency, 120 name-free answers, All 200 answers, Domain citations
Agency	120 name-free answers	All 200 answers	Domain citations
Omniscient Digital	15	49	41
Hamster Garage	3	19	39
Minuttia	1	15	24
Rampiq	0	14	25
Zupo	0	13	19

Five agencies accounted for 19 namings across 120 name-free answers. One firm took 15 of the 19.

Why we pool the four engines

The question people actually type is engine-specific. They want to know who wins in ChatGPT, or who wins in Perplexity. We are not going to answer that from this sample, and it is worth being direct about why.

Ten questions, five runs, four engines is 50 answers per engine. At that size the difference between an agency being named three times and being named once is inside the run-to-run variation we can already see in the pooled data. Publishing a per-engine league table from 50 answers would look more precise than the measurement is. A pooled count you can trust beats a split you cannot. If you need the per-engine split for your own category, that is what the AI Visibility Diagnostic produces, at 35 questions written for your category rather than a shared ten. It is $990 once, credited in full against the program if the client starts the program.

One thing did hold across all four engines: the same small group of names kept appearing, and the great majority of answers named at most two of them. The engines are not disagreeing about who exists. They are drawing from a narrow shared pool.

The citation layer

The 200 answers returned 663 citations across 223 distinct domains. 129 of those domains were cited exactly once. These counts describe source frequency in the recorded answers, not the impact of those sources.

Data table: Domain, Citations, Type
Domain	Citations	Type
beomniscient.com	41	Vendor
hamstergarage.com	39	Vendor
rampiq.agency	25	Vendor
minuttia.com	24	Vendor
discoveredlabs.com	20	Vendor
zupo.co	19	Vendor
yesoptimist.com	17	Vendor
embarque.io	15	Vendor

Every domain in the top group is owned by the agency it represents. The returned citations point to vendor sites.

Read the two tables together and the useful finding falls out. rampiq.agency was cited 25 times. Rampiq was named in zero of the 120 name-free answers. zupo.co was cited 19 times. Zupo was named in zero of them. Being cited is not being recommended. These observations do not measure leads or revenue.

What to do with this

If your own category behaves like this one, and most B2B software categories we have measured do, three things follow.

Vendor domains led the citation list in this category. That records source inclusion, not proof that every vendor was read. Compare citation and naming separately, and distinguish name-free questions from questions that already contained a name. The name-free count records unprompted naming in this sample; neither count predicts revenue from these data.

We publish the method in full: the exact questions, the engines, the scoring rules, the raw data and the limitations, so anyone can check the work or repeat it. The current measurement protocol is on the methodology page.

Method and limits

What we ran. Ten buyer questions, four engines, five runs, 200 answers, 26 to 27 July 2026, one run per prompt. Measured under method v1.0. Current engagements run under method v1.1, described on the methodology page. Question sets were regenerated for each run rather than held constant, so run-to-run movement is question variation rather than engine instability.

What we counted. An answer counts once for an agency it names, regardless of position or sentiment. Citations are URLs returned in an answer's source list, grouped by domain. They do not record engine visits.

What this does not prove. Nothing here measures the quality of any agency's work, its results, its pricing or its fit for your company. It measures how often four engines named each firm in that study, in one category over one window. A firm named zero times may be excellent and invisible. A firm named often may be neither.

Commercial disclosure. Broadcastwell publishes this page and provides the services described here. Research findings, editorial analysis and commercial claims are clearly separated.

Broadcastwell ran this measurement and was named in 0 of the 200 answers. It is not ranked here.

Related research source: The 2026 State of GEO, Volume I. Published open access under CC BY 4.0. Not peer reviewed. This disclosure applies to that source dataset; the agency measurement on this page is described separately above.

Questions

Which agencies do ChatGPT and Perplexity name for AI search visibility?

Pooled across Claude, ChatGPT, Perplexity and Google AI Overviews, the five agencies named most often across 200 answers were Omniscient Digital 49, Hamster Garage 19, Minuttia 15, Rampiq 14 and Zupo 13. On the 120 answers whose question named no agency, the same five were named 15, 3, 1, 0 and 0 times.

Why are the engine results pooled rather than split by engine?

Because the per-engine sample is too small to separate honestly. Ten questions run five times across four engines gives 200 answers in total, which is 50 per engine. At that size a difference of two or three names between engines is inside the run-to-run noise, so we report the pooled counts and do not publish a per-engine ranking.

Does being cited mean an engine recommends you?

No. Citation and recommendation are separate events. rampiq.agency was cited 25 times while Rampiq was named in zero of the 120 answers whose question named no agency. An answer can cite your domain while naming someone else.

Is Broadcastwell included in this ranking?

No. Broadcastwell publishes this page and provides AI search visibility services. The dated research findings are presented separately from current commercial claims and client evidence.

Start with a measured baseline

See which buyer questions leave you out, the sources visible in those answers, and the three changes worth testing first. The Category Audit covers ten questions; the Diagnostic extends to 35 questions and 525 scheduled observed answers, adaptive runs extra. AI answers vary, so repeated observations and disagreements are reported.

Start with the $490 Category Audit

Results appear in your account; a sign-in link arrives by email. Credits in full against the Diagnostic within 30 days.

Get the Diagnostic, $990

Results appear in your account; a sign-in link arrives by email. Credits in full against the program.

Every figure comes from the published method, v1.1.

Create a free account

Free 10-question check (one engine)