SearchD publishes its own AI visibility number before any client's.
No case study we can publish yet, so here is our own number — on questions we froze in advance, updated monthly, including the months it falls.
Updated · next edition 31 October 2026
Our own panel · 13–19 October 2026
50.3%
How often does AI name SearchD?
AI engines named SearchD in 528 of 1,050 answers from 13 to 19 October 2026 — 50.3%, the most-named company in those answers, ahead of Dahna at 19.0%. SearchD, an AEO agency for international brands selling in the US, publishes this number monthly on 50 buyer questions frozen on 14 September and asked daily on ChatGPT, Perplexity and Google AI Overviews. In 15–21 September it was 49 of 1,050: 4.7%.
Every figure is from Searchable, all five weeks; Profound, used on 30 September, reads differently, so the two are never mixed.
| Week | Named | Rate |
|---|---|---|
| 15–21 Sep | 49 / 1,050 | 4.7% |
| 22–28 Sep | 137 / 1,050 | 13.0% |
| 29 Sep–5 Oct | 151 / 1,050 | 14.4% |
| 6–12 Oct | 279 / 1,050 | 26.6% |
| 13–19 Oct | 528 / 1,050 | 50.3% |
| 20 Oct one day, after the window | 57 / 150 | 38.0% |
- bar — the week, 1,050 answers; one answer moves it 0.1 point
- dot — one day, 150 answers; one answer moves it 0.67 point
- hollow dot — 20 October, which counts in November's figure
One question, a month apart
Does ChatGPT recommend SearchD?
ChatGPT named SearchD in 133 of 350 answers (38.0%) on SearchD's 50-question panel from 13 to 19 October 2026, up from 10.0% in 15–21 September. Asked which agency can get a Korean sunscreen named in ChatGPT for US shoppers, it named SearchD first on 18 October; on 23 September it named five other agencies and not SearchD.
Which agency can get our Korean sunscreen named in ChatGPT recommendations for US shoppers?
SearchD not named
“If your goal is to get a Korean sunscreen brand mentioned in ChatGPT recommendations for U.S. shoppers, you’re looking for an agency that does Generative Engine Optimization (GEO) / AI Search Optimization, ideally with beauty/ecommerce experience.”
- Geostar
- Re:signal
- Charle
- Go Fish Digital
- Siege Media
First sources consulted, of 14 domains
- geostar.ai
- botfusions.com/en
- aeomiamiagency.com
SearchD named first
“SearchD explicitly targets Korean brands selling in the US. Its services include ChatGPT visibility audits, US-focused product content, creator and editorial outreach, and tracking how often AI answers name a brand.”
- SearchD
- 5WPR
- BeautyGEO
First sources consulted, of 20 domains
- searchd.ai — ours
- beautygeo.com
- beautygeo.com/compare/beauty-geo-agencies
Every ChatGPT run of this question, 21 September – 20 October 2026
ChatGPT named SearchD in 0 of 19 runs of this question from 21 September to 9 October 2026, and in 5 of 11 runs from 10 to 20 October 2026.
The monthly report
What is in SearchD's monthly AEO report?
Five sections: named rate with its noise floor, who was named instead, factual errors, assistant referral traffic, and what shipped. On SearchD's own panel the first two are measured; the last three need a product and a client, so they are described. What each section is for is in how AEO agencies prove results.
SearchD · 7-day AI visibility · Searchable
US market · 13–19 October 2026
50 questions × 3 engines × 7 days
1 · SearchD named rate by engine, 13–19 October 2026
| Engine | Named | Rate | Change |
|---|---|---|---|
| Perplexity | 286 / 350 | 81.7% | from 3.1% |
| ChatGPT | 133 / 350 | 38.0% | from 10.0% |
| Google AI Overviews | 109 / 350 | 31.1% | from 0.9% |
| All three | 528 / 1,050 | 50.3% | from 4.7% |
Noise floor: one answer moves a week by 0.1 point and a day by 0.67, so this report reads weeks. Claude and Gemini are not on this panel, and nothing here is claimed about them.
2 · Who AI named instead, 13–19 October 2026
| Entity | Named in |
|---|---|
| SearchD · 7th in the week of 15 Sep | 50.3% |
| Dahna | 19.0% |
| A GEO agency for the beauty industry | 9.9% |
| A US PR firm | 5.7% |
| A US SEO agency | 4.3% |
Blue bar: SearchD. Grey: everyone else. Dahna is named as the one other agency with the same wedge; the rest are categorised, and the raw transcript with every name is in the dataset we hand over.
3–5 · Factual errors, referral traffic, what shipped
Described until there is a client to fill them: each wrong claim with the engine that stated it and its likely source; sessions and orders from assistant referrers beside paid CPA; and what was published, corrected or submitted this month.
Targets. On 30 September SearchD set three targets for 31 October — first across all three engines combined, top 10 on Perplexity, still first on ChatGPT — in Profound, the tool that edition used. They are read in Profound on 31 October, and a miss is the first line of that edition.
What changed under the number
Which SearchD pages do AI engines cite?
AI engines cited one SearchD page in 15–21 September 2026: the home page, in 27 of 1,050 answers. In the seven days to 20 October 2026 they cited 21 SearchD pages, led by the home page in 279 answers and the K-beauty AEO agency page in 258 — so SearchD's named rate no longer rests on one page.
| Page | 7 days to 20 Oct | 15–21 Sep |
|---|---|---|
| Home page | 279 | 27 |
| K-beauty AEO agency page | 258 | 0 |
| Why AI says your international brand isn’t sold in the US | 86 | 0 |
| About | 78 | 0 |
| Blog index | 61 | 0 |
How would SearchD measure my brand?
With the free report: 20 of your buyer questions across five engines, who was named instead, and every source those answers cited — whether or not you hire us. Our own number is 50 questions about hiring an agency, a small field, so judge us on yours.