Module 01 · Scan

Your AI visibility baseline, in minutes.

One pass — site crawl, brand profile, four-engine sweep — so you know exactly where you stand before you spend a cent on content.

Typical run · 2–4 minutes
scan · acme-northbound.com
live
Resolving robots.txt600ms
Crawl · /pricing500ms
Crawl · /product500ms
Crawl · /blog/best-crm-2026600ms
Extract brand profile700ms
Probe ChatGPT · 12 prompts800ms
Probe Perplexity · 12 prompts800ms
Probe Gemini · 12 prompts800ms
Probe Claude · 12 prompts800ms
Compute baseline600ms
Visibility
0
Share of answer
0%
Citation coverage
0%
What the scan covers

Four passes. One baseline.

Every scan ships the same evidence bundle — so your team and your board are looking at the same numbers from day one.

Pass 01

Site crawl

Title, meta, H1/H2, markdown and navigation extracted from your priority pages. Robots-respecting, polite, capped.

Pass 02

Brand profile

Category, scope, target customers, use cases, themes, buyer problems and a starter competitor set — all derived from your real site.

Pass 03

Initial AI sweep

12 starter prompts across category, intent, comparison and citation buckets — run live against ChatGPT, Perplexity, Gemini and Claude.

Pass 04

Baseline metrics

Visibility Score, Share of Answer, Recommendation Rate and Citation Coverage — each with a confidence label.

Inside a scan

Numbers that ground the rest of the program.

Median runtime
0s
From button-press to baseline — for a standard 50-page domain.
Answers analysed
0
12 prompts × 4 engines — each parsed for brand, competitor and citation signals.
Source types
0
Every citation auto-classified into one of 11 deterministic source types.
Engines covered

All four. Every scan.

ChatGPT
ChatGPT
12 starter prompts · live
Perplexity
12 starter prompts · live
Gemini
Gemini
12 starter prompts · live
Claude
Claude
12 starter prompts · live
Why it matters

You can't fix what you can't see.

The scan replaces guesswork ("are we even cited?") with evidence: real answers, real sources, real competitors — captured the moment you sign up.

Every prompt links back to its verbatim AI answer.
Every metric carries a confidence label and sample size.
Every cited source is classified and scored on day one.
Every competitor mention is captured — even ones you didn't think to track.
Every scan is reproducible — re-run with the same prompts to compare baselines.
Time to first scan
< 3 min
From URL paste to profile and prompt suggestions.
Pages crawled
20–40
Firecrawl surfaces the pages that define your ICP and category.
Signals
12
Category, ICP, tone, competitors, geos, USPs and more.
Re-scan
Anytime
Trigger a fresh scan when you relaunch or reposition.
The scan pipeline
Step 01
Crawl

Firecrawl v2 pulls homepage, product and about pages.

Step 02
Extract

AI extracts category, ICP, competitors and tone signals.

Step 03
Confirm

You review and edit each signal — 60-second interstitial.

Step 04
Activate

Profile feeds prompt generation and scoring.

Do
  • Paste the primary marketing URL, not a landing page.
  • Review category and ICP — small edits unlock much better prompts.
  • Re-scan after any major messaging or product change.
  • Add differentiators the site doesn't spell out (e.g. 'UK-based').
Don't
  • Don't skip the review step — the scan is a draft, not a verdict.
  • Don't scan a redirect or gated page; the crawler will pick up thin data.
  • Don't run parallel scans on the same brand — the last one wins.
  • Don't over-narrow the ICP; wide-then-refine beats narrow-then-panic.

Keep reading

FAQ

What teams ask before they run a scan.

What does the scan actually do?+
Crawls priority pages on your domain, extracts category/audience/competitors, then runs a starter prompt set across all four major engines so you see your visibility baseline on day one.
How long does it take?+
Typically 2–4 minutes for a standard site. Larger plans crawl more pages.
What if my site blocks crawlers?+
We fall back to a manual setup tab where you paste content directly — and offer a CitationsBot/1.0 allowlist hint for your robots.txt.
What prompts are used in the starter sweep?+
Twelve prompts derived from your brand profile across category, buyer-intent, comparison and citation buckets — enough to ground a confident baseline.
Can I rescan?+
Standard plans include one monthly AI visibility scan. Extra scans consume remaining answer credits for the cycle.
Where is the data stored?+
Your sampled answers, prompts and citations are stored on EU infrastructure (UK + Frankfurt). See our Sub-processors page.
Try free

Run your first scan in 5 minutes.

5-day trial for $5. Your full baseline in the time it takes to grab coffee.