Skip to content
RookAtlas
On every plan, Free included

Know what AI says about you, and whether it's true

See how AI answers describe your brand, with the text behind each label, and compare claims with your team's fact sheet.

Sentiment requires an evaluated extractor to be enabled. Accuracy checks require a live checker. The dashboard states how many answers have been analyzed; missing analysis is never counted as neutral.

  • All eight AI engines on every paid prompt run
  • A 95% interval on visibility and citation rates
  • Every run priced before it starts

Opening soon: sign-up opens with live measurement, after our provider pilot.

How answers talk about you
Net sentiment+42over 113 mentions
Recommended36%41 of 113 mentions
Sentiment
  • Positive54
  • Neutral41
  • Mixed12
  • Negative6
Recommendation
  • Recommended41
  • Listed52
  • Alternative14
  • Cautioned6

RookAtlas tracks mentions and, when a live checker is enabled, compares claims with a fact sheet your team maintains. A separately enabled extractor adds sentiment, list positions and recommendation labels without changing the original mention measurements.

Measured on eight engines

  • ChatGPT
  • Gemini
  • Perplexity
  • AI Overviews
  • AI Mode
  • Claude
  • Grok
  • Muse
01

Tone you can count

The sentiment view counts positive, neutral, mixed and negative mentions. Net sentiment uses analyzed mentions and shows a prompt-based confidence interval when there is enough evidence.

  • Positive, neutral, mixed or negative per mention
  • Net sentiment from -100 to 100
  • The answers behind every label
Mentions, labelled
  • ClaudeP-02

    “JBL's Endurance Peak 4 is a strong pick for running: twist-lock ear hooks and IP68-rated earbuds.”

    PositiveRecommended
  • PerplexityP-01

    “The JBL Tour One M3 lasts up to 40 hours with noise cancelling on, though it costs more than JBL's Live and Tune pairs.”

    MixedListed
  • GrokP-04

    “JBL's Tune Beam 2 is rated only IP54, against IP68 for its Endurance Peak 4 earbuds.”

    NegativeCautioned
02

Claims checked against your facts

Keep up to 30 statements about your brand: prices, plans, where you operate. After each run, recent answers that name you are checked against them, and the accuracy rate comes with its interval.

  • Supported, contradicted or not covered
  • Accuracy rate with a 95% interval
  • A changed sheet starts fresh, never relabels
Your fact sheet, checked
Accuracy rate87%95% interval 74–94%
  • The Tour One M3 lasts up to 40 hours with noise cancelling on.14 supported3 contradicted
  • The Endurance Peak 4 earbuds are rated IP68.6 supported2 contradicted
  • The JBL Headphones app is on the App Store.11 supported0 contradicted
  • The Tune Beam 2 has adaptive noise cancelling.8 supported1 contradicted
03

Where a wrong claim comes from

A contradicted fact becomes an action listing the claims as the answers put them and the third-party sites those answers cite, so you know which page to correct or which editor to write to.

  • Distinct claims, quoted
  • The sites the answers relied on
  • Graded weak: the fix's effect isn't proven
A wrong claim and its sources
Wrong claimWeak

Your fact: The Tour One M3 lasts up to 40 hours with noise cancelling on.

As the answers put it
  • “The Tour One M3 lasts 70 hours with noise cancelling.” · 2 answers
  • “Around 60 hours with noise cancelling on.” · 1 answer
Third-party sites those answers citeAudio reviewsBudget audio guide

Contradicted in 3 of 17 checked answers.

How it works

  1. Step 1

    Write your fact sheet

    Up to 30 short statements your team stands behind.

  2. Step 2

    Run your prompts

    Mentions are recorded; a configured checker compares answers with your facts. The enabled sentiment extractor analyzes answers in the background.

  3. Step 3

    Correct the record

    Each wrong claim names the sources to fix or the editors to contact.

What's inside

  • Sentiment per mention

    Positive, neutral, mixed or negative, with the sentence it came from.

  • Net sentiment score

    Positive minus negative over mentions, from -100 to 100.

  • Recommendation labels

    Recommended, listed, alternative or cautioned.

  • Fact sheet

    Up to 30 statements per brand, versioned.

  • Accuracy rate

    Supported over supported plus contradicted, with its interval.

  • Wrong-claim actions

    Each with the claims and the sources behind them.

What it can't do

Checks judge answers against your sheet, not the model's knowledge of the world, so a claim no fact covers isn't judged. Correcting a page doesn't guarantee an engine changes what it says; the action is graded weak for that reason, while the contradiction itself is measured.

All features

Questions

Prices, credits and plans are on the pricing page .

How is sentiment measured?

The enabled extractor labels mentions positive, neutral, mixed or negative. Net sentiment is positive minus negative, divided by analyzed mentions, times 100. The view shows analysis coverage and a confidence interval; incomplete analysis is excluded.

What goes on a fact sheet?

Short statements your team is sure of: what you cost, which plans exist, where you operate, what you don't do. Up to 30 per brand. Editing the sheet starts a new version, and only checks against the current one count.

Why is a wrong claim graded weak?

The contradiction is measured, but no controlled study shows how often correcting a page changes what engines say. The grade describes the fix's evidence, not the finding.

Pick a plan. Every engine is included.

Sentiment and accuracy is on every plan, Free included. One credit runs a prompt on all eight engines (Free's five monthly credits run on ChatGPT, Gemini and AI Mode), and every run shows its price before it starts. Compare the plans and work out your credits.

Free tool, usable today: AI visibility checklist

Opening soon: sign-up opens with live measurement, after our provider pilot.

Plans from
$36/mo
Engines on every paid prompt run
8
Credits per prompt
1