An official research publication of ASN Intelligence LLC · Report ASN-R-2026-003 · September 2026
ASN Research › Measured datasets
Measured dataset · ASN Research

Which sources AI engines cite when someone asks for a funeral home

Across 4903 measured AI answers about funeral homes, 30% of citations came from directories, 67% from competing operators’ own websites, and 3% from editorial, forum or video sources. Measured by ASN Intelligence on 08 September 2026 across 655 engine runs.

Disclosure. ASN Intelligence provides paid services to Adams-Green Funeral Home, which appears in the tables below. We did not exclude them, adjust their figures, or rank them. The raw runs behind every number on this page are available on request.

What this measures, and what it does not

We asked web-grounded AI assistants the questions a family actually asks when choosing a funeral home, many times over, and recorded every source each answer drew on. This page reports where the citations come from. It is a level, measured on the date shown — not a change over time, and not a claim that any intervention caused anything.

A run that failed — an API error, an engine returning nothing — is discarded, never recorded as “not cited.” Counting a failure as a miss understates every figure, which is a mistake we made ourselves earlier this year and corrected publicly.

The headline finding. Deathcare has its own directory ecosystem, and it outranks the general ones. parting.com is cited more often than Yelp. So are dignitymemorial.com and funeralocity.com. A funeral home absent from those three is invisible to AI regardless of anything else it does — and most of the advice written about “AI visibility” never mentions them, because it was written for software companies.

What this measures, and what it does not

We asked web-grounded AI assistants the questions a customer actually asks in this category, many times over, and recorded every source each answer drew on. This page reports where the citations come from for funeral homes. It is a level, measured on the date shown — not a change over time, and not a claim that any intervention caused anything.

A run that failed — an API error, an engine returning nothing — is discarded, never recorded as “not cited.” Counting a failure as a miss understates every figure, which is a mistake we made ourselves and corrected publicly.

Directories and listing sites

These are claimable. For a business in this category they are the fastest available lever, because a listing can be created or corrected in hours rather than months.

SourceCitationsShare
parting.com2404.9%
dignitymemorial.com1633.3%
yelp.com1603.3%
checkbook.org1593.2%
funeralocity.com1402.9%
funeroso.com1002.0%
everloved.com941.9%
bbb.org901.8%
tributearchive.com681.4%
funeralus.com611.2%
legacy.com571.2%
checkbca.org410.8%

Operators’ own websites

These are competing businesses’ own sites. They cannot be claimed — they are useful because they show what a citable site looks like in this category, and because their combined share is the reason a business’s own website matters more than any single directory.

SourceCitationsShare
carewellcremations.com2505.1%
adamsgreen.com2455.0%
adamsgreen.funeraltechweb.com1553.2%
bergepappassmith.com1162.4%
turrentinejacksonmorrow.com1152.3%
dfsmemorials.com901.8%
ekoji.org901.8%
herndon-va.gov851.7%
nirvanaagency.com711.4%
virginiacremate.com651.3%
rfhr.com641.3%
minhdangquang.org511.0%

Editorial, forum and video sources

Earned rather than claimed. In categories where directories are a small share, this is the only route.

SourceCitationsShare
facebook.com891.8%
reddit.com140.3%
linkedin.com140.3%
youtube.com80.2%
en.wikipedia.org40.1%

Method

Limits

Engines change, and a model update can move these shares without anything changing on the open web. This is one operator's measurement, published so it can be checked, not an industry consensus. Figures without a sample size beside them should not be quoted — including ours.

Want the raw runs? Email contact@asnintelligencellc.com and we will send the underlying data. We publish our method and our null results at ASN Research.