Tools for finding and evaluating search keywords and topics. Ranked by how often ChatGPT, Perplexity, Gemini, Copilot, Google AI Mode mention and cite each brand when asked 5 buyer questions.
Snapshot for the week of August 17, 2026. Named = the engine wrote the brand into its answer. Cited = the engine linked the brand's own domain as a source, out of 25 answers.
How often each engine named the brand, as a share of the questions it answered. Reading across a row shows whether the engines are describing the same market — the spread column is the gap between the engine that names the brand most and the one that names it least.
| ChatGPT | Perplexity | Gemini | Copilot | Google AI Mode | Spread | |
|---|---|---|---|---|---|---|
| Semrush | 100 | 80 | 100 | 100 | 100 | 20 |
| Ahrefs | 100 | 80 | 100 | 100 | 100 | 20 |
| KWFinder | 40 | 60 | 60 | 80 | 40 | 40 |
| AnswerThePublic | 20 | 40 | 80 | 60 | 80 | 60 |
| LowFruits | 40 | 0 | 80 | 20 | 20 | 80 |
| Ubersuggest | 20 | 0 | 0 | 60 | 60 | 60 |
| KeySearch | 0 | 0 | 40 | 0 | 20 | 40 |
| Keywords Everywhere | 20 | 20 | 0 | 20 | 0 | 20 |
| Wordtracker | 0 | 20 | 0 | 0 | 0 | 20 |
Values are the percentage of that engine's answers naming the brand. Each engine answered the same 5 questions here, so a cell moves in steps of 20 points. 2 brands are named in more than half of one engine's answers and never mentioned by another — the engines are not describing the same shortlist.
Two different kinds of visibility. The filled dot is how often engines wrote the brand into the answer; the hollow dot is how often they linked the brand's own domain as a source. The bar between them is the gap — and where the filled dot sits to the left, the brand is cited more often than it is recommended.
Gap column is percentage points, named minus cited. Widest this week: Ahrefs, named in 96% of answers but cited in 28%.
Domains the engines cited most across this category's answers this week. These are the surfaces that shape AI recommendations here, and where coverage is worth checking first.
Every week cloro asks each engine the same 5 buyer-intent questions with US localization, then scores the answers with deterministic string and URL matching — no LLM judge, so the same answers always give the same score. The 0–100 AI Visibility Score weights mention rate (60%), citation rate (25%), and how early a brand is named (15%).
Read the full methodology →cloro. (August 17, 2026). AI Visibility Leaderboard — Keyword Research Tools (Week of 2026-08-17) [Data set]. cloro Research. https://cloro.dev/ai-visibility/keyword-research/
Underlying data · CC BY 4.0
Week of 2026-08-17, unchanged since publication. The page above refreshes weekly.
cloro.dev/ai-visibility/data/2026-08-17.json
@misc{cloro_ai_visibility_leaderboard_keyword_research_tools_2026_08_17,
title = {AI Visibility Leaderboard — Keyword Research Tools (Week of 2026-08-17)},
author = {{cloro}},
year = {2026},
howpublished = {\url{https://cloro.dev/ai-visibility/keyword-research/}},
note = {Data set. cloro Research. Snapshot: \url{https://cloro.dev/ai-visibility/data/2026-08-17.json}}
}The score runs 0–100 and combines three signals: how often the brand is mentioned across engine answers (60%), how often an engine cites the brand's own domain as a source (25%), and how early the brand appears when it is named (15%). Detection is deterministic string and URL matching, not an LLM judgement, so the same answers always produce the same score.
This snapshot covers ChatGPT, Perplexity, Gemini, Copilot, Google AI Mode. Every engine is asked the same 5 buyer-intent questions with US localization, giving 25 engine answers for this category.
Weekly. A fresh sweep runs every Monday (UTC) and each page is stamped with the week it reflects. Rankings are deliberately not real-time — AI recommendations do not move hour to hour, and dated snapshots are easier to cite and compare.
A mention means the engine named the brand in its answer text. A citation means the engine linked the brand's own domain as a source. They are tracked separately because they are different kinds of visibility — a brand can be cited as a source without ever being recommended, and vice versa.
Yes. Everything here is free to reuse with attribution under CC BY 4.0 — cite the category page and the week it covers.
The leaderboard runs on the same API that monitors your brand across ChatGPT, Perplexity, Gemini, Copilot, and Google AI Mode. Start with 500 free credits a month.