CiteGround teardown / citeground-category

When your buyers ask AI about your category, who gets named?

citeground.com · 12 buyer prompts · 118 runs across ChatGPT (API, web-grounded), Google AI Overviews, Perplexity (API) · rolling 28-day window · generated 2026-07-03 16:02 UTC

Our own tool, day 0 of a public build: the pre-launch baseline in our own category. We publish the gap before the work.

CiteGround is absent from the AI answer on 12 of 12 buyer prompts. Every number below carries its run-to-run variance band: AI answers are noisy, and a single run is not evidence. We show the noise instead of hiding it.

Share of voice across the basket

Fig. 01 / share of answers mentioning the brand118 runs

Profound28%

Otterly26%

Peec15%

Scrunch14%

Trakkr8%

Frase6%

CiteGround (us)0%

Share of all 118 archived runs in the window that mention the brand. Basket-level only; per-query AI rank is statistical noise and is never reported here.

Prompt-by-prompt: what the AI actually says

Prompt 0112 runs

“best AI visibility checker for my brand”

CiteGround absent (0/12 runs)set agreement 33% (unstable)

The AI names:

Otterly (5/12)Frase (3/12)Profound (3/12)Peec (1/12)

Grounding sources: youtube.com, semrush.com, ahrefs.com (35 distinct domains cited across runs)

Prompt 0212 runs

“how do I check if ChatGPT recommends my product”

CiteGround absent (0/12 runs)set agreement 100% (stable)

Grounding sources: openai.com, reddit.com, bazaarvoice.com (18 distinct domains cited across runs)

Prompt 0312 runs

“free tool to see which brands AI answers mention”

CiteGround absent (0/12 runs)set agreement 83% (mixed)

The AI names:

Peec (1/12)

AI answer returned in 9/12 runs.

Grounding sources: visibrand.app, areyoumention.com, viametric.app (36 distinct domains cited across runs)

Prompt 0410 runs

“track my brand mentions in ChatGPT and Perplexity answers”

CiteGround absent (0/10 runs)set agreement 33% (unstable)

The AI names:

Otterly (3/10)Peec (3/10)Profound (2/10)

Grounding sources: youtube.com, keyword.com, adobe.com (27 distinct domains cited across runs)

Prompt 059 runs

“best AEO tools for B2B SaaS”

CiteGround absent (0/9 runs)set agreement 67% (mixed)

The AI names:

Profound (5/9)Otterly (2/9)

Grounding sources: hubspot.com, youtube.com, reddit.com (38 distinct domains cited across runs)

Prompt 069 runs

“Otterly alternative for AI search monitoring”

CiteGround absent (0/9 runs)set agreement 67% (mixed)

The AI names:

Otterly (9/9)Peec (5/9)Scrunch (3/9)Profound (3/9)Trakkr (1/9)

Grounding sources: otterly.ai, ziptie.dev, nightwatch.io (24 distinct domains cited across runs)

Prompt 079 runs

“Profound alternative for startups”

CiteGround absent (0/9 runs)set agreement 33% (unstable)

The AI names:

Profound (9/9)Trakkr (5/9)Otterly (4/9)Scrunch (4/9)Peec (3/9)

Grounding sources: trakkr.ai, seranking.com, writesonic.com (22 distinct domains cited across runs)

Prompt 089 runs

“how to measure share of voice in AI answers”

CiteGround absent (0/9 runs)set agreement 100% (stable)

The AI names:

Trakkr (3/9)Scrunch (3/9)

Grounding sources: youtube.com, netranks.ai, semrush.com (20 distinct domains cited across runs)

Prompt 099 runs

“tool that shows which sources AI engines cite”

CiteGround absent (0/9 runs)set agreement 33% (unstable)

The AI names:

Profound (4/9)Otterly (3/9)Scrunch (3/9)Peec (2/9)

AI answer returned in 8/9 runs.

Grounding sources: therankmasters.com, siftly.ai, otterly.ai (38 distinct domains cited across runs)

Prompt 109 runs

“check my website AI search visibility”

CiteGround absent (0/9 runs)set agreement 100% (stable)

AI answer returned in 8/9 runs.

Grounding sources: semrush.com, webaloha.co, seranking.com (14 distinct domains cited across runs)

Prompt 119 runs

“generative engine optimization tools comparison”

CiteGround absent (0/9 runs)set agreement 33% (unstable)

The AI names:

Profound (6/9)Frase (4/9)Otterly (3/9)Peec (3/9)Scrunch (3/9)

Grounding sources: youtube.com, tryprofound.com, yotpo.com (39 distinct domains cited across runs)

Prompt 129 runs

“AI search rank tracker that is actually reliable”

CiteGround absent (0/9 runs)set agreement 67% (mixed)

The AI names:

Otterly (2/9)Scrunch (1/9)Profound (1/9)

Grounding sources: rankscale.ai, reddit.com, nightwatch.io (41 distinct domains cited across runs)

See the move sequence to get CiteGround into these answers

Method: each prompt runs multiple times per engine; brands and cited domains are extracted per run and archived append-only from run #1. Counts of runs, not certainty theater. Presence counts (k/n) and set agreement are two different measurements: an answer can include a brand in every run while the full named set still churns. Engine access, reading construction, and what we exclude: /methods.