Introducing Sight: one score for SEO and AI visibility

Your website now has two audiences: search engine crawlers and AI answer engines. Google still sends the bulk of your traffic, but a growing share of buyers start their research in ChatGPT, Claude, or Perplexity instead of a search box. Those tools don't rank pages — they read them, summarize them, and decide whether to cite them. Sight is the audit we built because no existing tool measured both jobs at once, and treating them as one problem turns out to matter.

The problem: two audiences, half a picture

Most audit tools only look at one side. They'll flag a missing meta description or a slow Largest Contentful Paint, but they have nothing to say about whether an AI answer engine can actually parse your content, trust your claims, or find a clean quote to cite. Meanwhile the handful of AI-visibility tools that have appeared tend to ignore the classic-SEO fundamentals that AI crawlers still depend on.

That split is a false one. The technical foundation is shared: an AI crawler reaches your site the same way a search crawler does, reads the same HTML, and benefits from the same clean structure and fast pages. A page that's blocked in robots.txt, hidden behind a client-side render, or buried under vague content fails both audiences at once. Running two separate tools to find that out — and reconciling two conflicting punch lists — is busywork. So we didn't.

What Sight does

Sight scans your site once and scores both sides of the equation in a single pass. You get one number and one prioritized list of fixes instead of two reports that don't talk to each other.

On the classic-SEO side, it checks the things that have always mattered: the meta tags that carry real weight — title, description, canonical, Open Graph — plus indexability, structured data, links, mobile, security, and Core Web Vitals measured page by page.

On the GEO side, it checks what AI answer engines need: whether the AI assistant crawlers can reach you, whether your content is readable without JavaScript, whether your structure and markup make claims easy to extract, and how often models actually mention you versus your competitors when asked questions in your space.

Every finding lands in one report, ranked worst-first — the highest-severity failure at the top, so you always know what to fix next.

Why one score

Because a single number forces honesty. Two dashboards let you feel good about a green SEO score while quietly failing every AI-visibility check — or the reverse. One blended score, built from clearly weighted categories, tells you where you actually stand for the whole web as it exists now, not the half of it your old tool measured. And one punch list means you fix the thing that helps both audiences first, instead of optimizing a title tag on a page that's invisible to AI crawlers anyway.

That's the whole idea behind the name: one clear look at your site the way both a search engine and an answer engine see it — and one honest, fixable score. If you want the deeper argument for why the AI side is now worth measuring, our post on what actually changed between SEO and GEO lays it out.

What a scan gives you

Point Sight at a URL and it crawls your site, measures both the classic-SEO and GEO sides, and returns a scored report you can act on immediately. Each finding carries a severity and a concrete fix, and the whole list is ordered so the highest-impact problem sits at the top — no triage required on your end. The most common surprise, over and over, is the gap between how a site looks in a browser and what a crawler that doesn't run JavaScript actually receives. If that's you, our post on why your site might be invisible to AI explains exactly how it happens and how to fix it.