AI Visibility

See how AI talks
about your brand

Buyers ask ChatGPT and Perplexity before they ask you. Track whether your brand shows up across eight AI models, how it's described, and who's winning the conversation.

Tracked across eight AI models, scanned daily

ChatGPT
Gemini
Perplexity
Claude
Google AI Overviews
Google AI Mode
Grok
DeepSeek

A full picture of your AI presence

Everything you need to monitor, benchmark, and improve how AI represents your brand.

8 models, daily scans

ChatGPT, Gemini, Perplexity, Claude, Google AI Overviews, Google AI Mode, Grok and DeepSeek — automatically scanned every day.

Core visibility metrics

Visibility %, Share of Voice, Sentiment, and Average Position, tracked over time and broken down by model.

Prompt tracking

Group prompts by Discovery, Brand, and Competitor intent, then see per-prompt performance across every model.

Slice by intent, branded & features

Every prompt is auto-classified by buyer intent (Commercial / Informational / Navigational) and deterministically as branded vs non-branded, so you can slice visibility in one click — and each slice shows its sample size. Filter answers by what the engine actually did, like whether it ran a live web search.

Competitor benchmarking

Line your brand up against rivals on a competitive quadrant and watch share of voice shift week to week.

Sources & citations

See exactly which domains and pages AI cites when it answers — the content actually shaping the response.

Insights & actions

Prioritised, plain-language recommendations on where to win mentions and close the gaps competitors are filling.

Shopping visibility

Track whether AI recommends your brand — and which of your products — for purchase-intent shopping questions, benchmarked against competitors and tied to your imported catalog.

Crawlability & crawl insights

Check which AI crawlers your robots.txt allows — with why and how to fix it — then forward your access logs to see which AI bots (GPTBot, ClaudeBot, PerplexityBot and more) actually visit, how often, and which pages they hit.

GEO Audit + Draft the fix

A 0–100 AI-readiness score across your whole site — crawler access, server-rendered content, structured data and llms.txt — with a prioritised fix list. Let Crunch draft a copy-pasteable fix for any issue.

Import from Peec AI

Already track AI visibility in Peec AI? Connect a project with an API key to pull its visibility, sentiment, prompts and sources straight into your reports.

Track

Eight models, scanned every day

CrunchJunkie asks the questions your customers ask and records how each model answers — whether your brand shows up, where it ranks, and how it's described. No manual prompting, no copy-paste.

  • ChatGPT, Gemini, Perplexity, Claude, Google AI Overviews, Google AI Mode, Grok, DeepSeek
  • Fully automated daily scans
  • Mention tracking with full response capture
  • Trend lines per model over time

Visibility by model

Daily scan
ChatGPT
71%
Perplexity
64%
Gemini
58%
Claude
55%
Google AI Overviews
49%
Google AI Mode
46%

Lumière — visibility overview

Across 8 models · last 30 days

Visibility

62%+8 pts

Share of Voice

31%+4 pts

Avg. Position

2.4−0.6

Sentiment

+0.71Positive
Measure

The metrics that tell you if you're winning

Visibility tells you how often you appear. Share of Voice tells you how much of the conversation you own. Sentiment and Average Position tell you whether that presence is working for you.

  • Visibility % across all tracked prompts
  • Share of Voice versus competitors
  • Sentiment scoring on every mention
  • Average position when you do appear
Benchmark

See where you stand against rivals

Add your competitors once and CrunchJunkie tracks them on the same prompts you do. The competitive quadrant shows who owns the conversation and who's gaining — so you know where to push.

  • Side-by-side visibility and share of voice
  • Competitive quadrant view
  • Per-prompt win and loss tracking
  • Mention-gap detection where rivals appear and you don't
BrandVisibilityShare of Voice
LU
Lumière (you)
62%31%
VE
Velora
58%27%
NO
Nortex
44%19%
BR
Brandfield
38%15%

Most-cited sources

Domains AI references when answering

g2.com
41 cites
reddit.com
33 cites
trustpilot.com
27 cites
capterra.com
19 cites
wikipedia.org
12 cites
Understand

Know which sources shape the answer

AI answers are built from sources. CrunchJunkie surfaces the domains and pages each model cites most, so you can see whether it's review sites, forums, or your own content driving how you're described.

  • Ranked list of most-cited domains
  • Page-level citation detail
  • Spot which content earns mentions
  • Find the gaps to target with new content
Prompts

Track the questions that matter, by intent

Organise prompts into three buckets so you know whether you're winning discovery, owning your brand, or losing comparisons.

Discovery

Category questions where buyers haven't named a brand yet — "best marketing reporting tools for agencies". Winning here means being recommended before anyone asks for you.

Brand

Prompts that name you directly — "is Lumière any good?". This is where sentiment and accuracy matter most, because the model is describing you to a buyer with intent.

Competitor

Comparison and alternative prompts — "Lumière vs Velora", "alternatives to Nortex". Track whether you show up favourably when buyers are weighing options.

New · AI Shopping Visibility

Does AI recommend your products for the sale?

Buyers ask AI what to buy. Track your shopping recommendation rate against competitors, detect which of your products AI already suggests, and import your catalog by feed or Google Merchant Center — all inside AI Visibility.

Explore

Honest numbers, with the sample size shown

AI answers are non-deterministic, so a visibility figure is a sample — and we treat it like one. Every AI-visibility KPI shows how it was measured, in the app and in your client reports alike.

Sample size on every metric

Each KPI shows "Based on N runs" — the number of scan runs (prompts × models × repeats) behind the number — so nobody mistakes one lucky answer for a trend.

A margin of error, not false precision

Visibility % is shown with a ± margin of error (its standard error), so a client can see the confidence around the figure at a glance.

Repeated runs, averaged

We run each prompt several times and aggregate, because a single manual check is one noisy sample. The trend you track is real, not a screenshot.

Want to see it first? Run the free AI-visibility check — a real sample with its sample size and margin of error, no signup.

Bring your own keys, at cost

Already have API keys for the models you care about? Connect them and you only ever pay the providers their usage rates — CrunchJunkie never marks up your AI spend. Keys are encrypted at rest and used solely to run your scans.

Or let us manage it — metered from €25/mo

Don't want to juggle API keys? Managed AI runs the scans for you on our infrastructure. Model choice is free — you pick a monthly plan with an included check allowance (from €25/mo) and pay just €0.05/check beyond it. Premium models count 3×. Turn models on and off whenever you like — set a spend cap so there are no surprises.

See Managed AI pricing

Find out what AI says about you

Run the free check to see a real, live sample in under a minute — no signup — then set up a full visibility project when you're ready.

Free check needs no signup · 14-day trial · Bring your own keys or go managed