From a website to a clear picture
- 1
Find the company's website
We resolve the address you enter, follow redirects and read robots.txt, the sitemap and llms.txt. Sites that block automated visitors are reported as such; we never guess.
- 2
Read it like a first-time visitor
We read the homepage first, then the pages that describe the company: about, services or products, pricing, contact, team. Text, structured data and the site's code (fonts, colors, technologies) are extracted.
- 3
Turn observations into evidence
Every conclusion is a claim backed by an exact excerpt from a source. The company's own statements are kept apart from what we observe and from what we infer.
- 4
Let specialists disagree
Nine specialist agents look at clarity, brand, user experience, conversion, search, AI discoverability, trust and growth. A skeptic challenges conclusions the evidence can't carry, an evidence agent checks every reading, and a judge decides how confident each conclusion is. Disagreements are shown, not hidden.
- 5
Score only what can be observed
Scores are calculated from signals with a fixed formula: each signal's value, its importance, how reliably it was observed, and whether it matters for this kind of company. The same input always gives the same score.
- 6
Prioritize what to fix
Recommendations state the problem, the evidence, the action, and the priority. Priority weighs impact, visibility and urgency against effort.
What the evidence labels mean
Every fact shows how it is known. Missing information is shown as missing, never filled in. A company without a stated mission gets “Not explicitly found”, not an invented one.
- Official. The company states this itself.
- Observed. Seen directly on the website or in a source.
- Inferred. Our reasoning from the evidence. Not stated by the company.
- Unknown. Not enough evidence to say.
- Not found. We looked and couldn't find it. That doesn't mean it doesn't exist.
What the reports cover
The free snapshot shows what matters most. Reports investigate one side of the company in more depth.
- Brand intelligence
- Positioning, identity, consistency and visual identity, with evidence.
- Digital presence
- Expected versus actual information, and what is buried.
- AI visibility readiness
- Signals that help AI assistants identify and describe a company. Readiness, not measured visibility.
- Website conversion
- How clearly the site turns interest into action, and what it measures.
What it doesn't do
- It doesn't measure how often AI assistants mention a company. It measures the signals that make that possible.
- It doesn't claim to know how the market perceives a company. It reports how the website presents it, and what independent sources say when they can be verified.
- It reads public pages only, and respects sites that ask crawlers to stay out.