How it works

Paste a URL. Get the truth. Fix it.

There's no tag to install, no data warehouse to connect and no onboarding call standing between you and your first score. Here's exactly what happens, in order.

You give us a domain

That's the whole setup. We crawl the site the way an AI engine's fetcher does — same user agents, same rendering constraints, same patience for slow servers. On a typical marketing site the first pass finishes in under two minutes.

No tag to install No analytics access No card

We score 80+ readiness signals

Everything a model needs in order to find you, parse you and trust you. Signals roll up into four sub-scores and one composite number, and every failure is listed with the URLs that caused it.

Crawl access

Can each AI fetcher reach the page at all?

Structured data

Is the machine-readable layer present and valid?

Content extraction

Can a clean answer be lifted out of the page?

Trust signals

Does the rest of the web corroborate what you claim?

We ask the engines about you

You start with a prompt set seeded from your category and your competitors, then you edit it — the generic, brand-heavy prompt libraries other tools ship produce a flattering number nobody can act on. Each prompt runs repeatedly across the month, because answer engines are generative and a single reading is a sample, not a rank.

What a run looks like

engines queried

30×

repeats per prompt, per month

confidence range on every figure

You get a ranked fix list

Issues sorted by impact against effort, each with the artefact that closes it: validated JSON-LD, a content brief written against the prompts you're losing, or a short checklist. Ship, re-scan, and the history shows whether it landed.

01Add Organization + FAQPage JSON-LD to 14 pages+14 PTS

02Restructure 6 service pages answer-first+9 PTS

03Resolve entity name mismatch across 3 directories+6 PTS

How long before anything moves?

Four to eight weeks, honestly. Engines have to re-crawl your pages and the models have to start drawing on them. Some fixes register in days — a robots.txt rule that was blocking a fetcher, a broken canonical. Entity and schema work takes longer, because it depends on other sites catching up too. Anyone promising a jump in a fortnight is selling you something.

DAYS 1–7

Access and technical fixes register. Crawl score usually moves first.

WEEKS 2–4

Schema and restructured pages start showing up in retrieval.

WEEKS 4–8

Citations follow. This is where the visibility line usually turns.

Methodology

Check our numbers yourself

How do I know the scores are accurate?

Spot-check us. Run the free audit, then run five of the same prompts by hand in ChatGPT, Perplexity and Google AI Mode and compare. That's the standard advice for buying anything in this category, and we'd rather you did it on us than on a competitor. Our methodology is published in full: how many prompts per engine, how often, from which locations, and what counts as a mention.

Why does my score move when I've changed nothing?

Because answer engines are generative, not deterministic. The same prompt can return a different answer twice in a row. We run each prompt repeatedly across the month and report the pattern with a confidence range attached, rather than one number pulled from one run. It's also why we show trend lines by default and treat a single day's movement as noise until it holds.

Where do the prompts come from — can I use my own?

Yes, and you should. We seed you with a set built from your category and competitors, then you edit it. If your buyers ask something specific, track that instead of a generic industry phrase.

Does this replace my SEO tooling?

No, it sits alongside it. Your rank tracker measures a deterministic results page; this measures which brands get named and cited inside a generated answer. The two overlap on technical foundations and diverge everywhere else — different prompts, different competitors, and often a different set of pages doing the work.

What happens to my data?

Scan data is processed in the EU and US and we're GDPR-ready. We don't sell it, pool it into a public index, or use your prompt sets to train anything.

Two minutes from now you'll know where you stand

Free scan, no card, no sales call. Keep the report either way.

Scan your site free