AI search news ·
What sits around a first AI citation: all 60 slots Perplexity filled on the question that finally cited us — 25 distinct pages, zero head-to-heads
On 2026-08-10, Perplexity cited askedabout.com at position 3 of 20 sources in 3 of 3 identical samples of “What is a good Profound alternative for a small business?” — the first non-zero cell in our weekly self-citation gauge's history (run 1, 2026-08-03: 0 of 144). Instead of celebrating, we dumped the company it keeps: all 60 citation slots Perplexity filled across the three samples. They contain 25 distinct pages from 24 domains — 21 multi-vendor “alternatives” lists, one directory alternatives page, two Reddit threads asking for alternatives, one single-vendor pitch — and not one head-to-head “X vs Y” comparison. The page it cited is the one list-shaped comparison we have ever shipped, built 2026-07-31 to test exactly this shape hypothesis. Ten days later the question it targets is the one question that cites us.
Primary source: AskedAbout — self-citation gauge run 2 (2026-08-10, 144/144 production answers), sample-level citation payloads read verbatim
Method, stated before the numbers
The gauge is the same fixed instrument as our weekly self-audit: 12 pinned small-business buyer questions × 4 engines (ChatGPT, Perplexity, Gemini, Claude) × 3 identical samples = 144 production answers per run, executed on the exact engine that powers a customer's audit. A citation means an askedabout.com URL appears in the engine's returned citation payload for that answer, domains normalized hostname-minus-www — the same rule as every dated cut in the citation-economy series. Two dated runs exist: 2026-08-03 and 2026-08-10. This post publishes the raw cells of the only non-zero reading, plus every URL cited alongside us.
The gauge history: 288 cells, 285 zeros, one repeated non-zero
| Run | Date | Answers | Cells citing askedabout.com | Per-engine split |
|---|---|---|---|---|
| Run 1 | 2026-08-03 | 144 (0 failed) | 0 | ChatGPT 0/36 · Perplexity 0/36 · Gemini 0/36 · Claude 0/36 |
| Run 2 | 2026-08-10 | 144 (0 failed) | 3 (2.1%) | ChatGPT 0/36 · Perplexity 3/36 · Gemini 0/36 · Claude 0/36 |
All three non-zero cells are the same (engine, question) pair: Perplexity × “What is a good Profound alternative for a small business?”. The other 141 answers in run 2 cite us nowhere, and the mention rate — an engine actually writing the name AskedAbout in its answer text — was 0 of 144 in both runs. Perplexity linked our page three times and never once said our name.
The three cells, verbatim
| Sample | Sources in citation payload | askedabout.com position | Cited | Mentioned by name |
|---|---|---|---|---|
| 1 of 3 | 20 | 3rd | yes — /compare/best-profound-alternatives | no |
| 2 of 3 | 20 | 3rd | yes — same URL | no |
| 3 of 3 | 20 | 3rd | yes — same URL | no |
Position 3 of 20, three times in a row, behind scoutrival.com (position 1 in all three samples) and ahead of seranking.com, birdeye.com, and writesonic.com. By our own within-run stability data, only 28.9% of citations appear in 3 of 3 identical asks — the new citation entered at the stable end of the distribution.
The co-citation census: what shape gets cited on this question
| Shape of co-cited page (by its own URL and slug) | Distinct pages | Share of 25 |
|---|---|---|
| Multi-vendor “alternatives” list (blog/guide/resource page) | 21 (ours included) | 84% |
| Directory alternatives page (g2.com) | 1 | 4% |
| Reddit thread asking for alternatives | 2 | 8% |
| Single-vendor pitch page | 1 | 4% |
| Head-to-head “X vs Y” comparison | 0 | 0% |
60 citation slots, 25 distinct URLs, 24 domains; 15 domains appear in all three samples. Not one head-to-head comparison appears in any slot — no URL in the entire set contains a “vs”. That asymmetry is why our page exists: when we read this question's citation set on 2026-07-28, every cited page was a multi-vendor list while every comparison page we had live was a head-to-head. We shipped one honest multi-vendor list — 18 tools, every entry price read from the vendor's own pricing page on 2026-07-31, our own tool listed with its real limits — and pre-registered the question it targets. Ten days later the gauge read non-zero on exactly that question, on exactly that page, and on no other cell.
What this is not, stated plainly
One repeated cell is direction, not victory. Three of 288 lifetime gauge cells are non-zero; ChatGPT, Gemini, and Claude still cite us nowhere on the panel; the mention rate is still zero; and our churn data says a domain cited by a single engine survives the next weekly run 49.0% of the time — a coin flip. The falsification is already scheduled: run 3 fires 2026-08-17, and we publish the survival read either way. If the citation is gone, the shape hypothesis loses most of its first supporting datum and we will say so in this series.
The takeaway if you are chasing AI citations
Before writing the page you hope an engine will cite, read what the engine already cites on that exact question. If the citation set is 100% lists — as it was here, 25 of 25 pages, zero head-to-heads — a head-to-head cannot join it by being better; it is the wrong artifact shape for the slot. That is a checkable, per-question fact, not a general theory: our engine reads the citation payload for any business's buyer questions the same way, and the free 60-second check shows you the source list the engines actually return.
What finally got a page cited by an AI engine after 285 zero readings?
Matching the artifact shape of the existing citation set. Every page Perplexity cites on the target question is a multi-vendor alternatives list, a directory page, or a community thread — never a head-to-head comparison. The first page we shipped in that list shape (2026-07-31, 18 vendors, dated prices) was cited at position 3 of 20 in 3 of 3 identical samples ten days later, while all our head-to-head pages remained uncited on the panel.
Does an AI citation mean the engine recommends you by name?
No. In all three citing samples, Perplexity attached our URL to its citation payload but never wrote the name AskedAbout in the answer text — cited, not mentioned. Citation and mention are separate measurements: the gauge's mention rate stayed 0 of 144 in the same run.
How likely is a first-week AI citation to survive?
By our own published base rates: 49.0% week-over-week survival for domains cited by a single engine (measured twice, on 520 and 538 domains), and only 28.9% of citations appear in 3 of 3 identical samples. Our new citation shows the stable 3-of-3 pattern but is single-engine. The scheduled 2026-08-17 re-run is the test, and we publish the result either way.
See your number
See which businesses AI names when your client's buyers ask.
Running this for clients? The $249 agency 5-pack audits five businesses, white-labeled.
Who runs this
- Built and operated by Sensara LLC, Atlanta, Georgia — about us and how the audit works.
- See what the report looks like before you run anything — score per engine, the competitors AI names instead of you, and a fix plan.
- We run the same audit on ourselves every week and publish the result: in the latest run AI named AskedAbout in 1 of 144 answers. We report our own numbers the way we report yours.