The site audit slides

Read the site audit block to see whether AI assistants can reach, read and quote your website at all.

Site audit slides overview

This is the first block carrying your own data, and it answers a question that comes before every other one in the report: can an assistant use your site?

It appears only once your brand has a completed site audit. Everything here mirrors the Site screen in the app, where the same findings are live and clickable — see Site audit.

There are up to eight slides. The last four are about your homepage specifically, because on a first audit that’s the page that gets the deep treatment.

Overall site audit

The domain summary, with the three scores side by side, plus whether AI bots are allowed in.

SRPS is how ready your pages are for assistants to find, read and cite. SRPS+ is a separate, stricter score that exists only where a deep audit ran. PERF is the speed measurement.

Each is out of 100, green from 80 up (90 for PERF), amber in the middle, red below. The target is 100 — every point missing is a check that didn’t pass.

For what each one measures and doesn’t, read SRPS, SRPS+ and PERF.

Good to know A score of 0 is almost never a hundred small failures. It normally means one blocking problem — the page errors, blocks AI crawlers, or is marked “don’t index” — while everything underneath is fine. Fix that one thing and the real score appears.

Indexability, canonicalization and entity

Three related questions on one slide:

• Can your pages be indexed and discovered? If not, they can’t be cited.
• Do you have one canonical address per page? When the same content answers at several addresses, the value is split and neither version wins.
• Does your site state clearly who you are? The brand-entity markup that tells a search engine or a model that this site belongs to a specific, identifiable organisation.

The third is the one people skip and the one that matters most for AI answers: an assistant that can’t tell who you are has nothing to attach a recommendation to.

Site structure by page

Every page the crawl found, arranged as a tree: languages at the top level (read from your hreflang tags), then the real sections of your URLs — /blog, /products, and so on.

The subtitle tells you how many pages across how many locales.

Read it for shape, not detail. Sections you expected and don’t see, or a language branch that’s much thinner than the others, are the findings here.

Broken links across the site

Links that returned an error, collected across every page the crawler checked. Internal links come first, because those are entirely yours to fix.

Each row shows the link text, and which page it was found on.

Where the list is truncated, the slide says how many more there are — the complete list is in the app.

Some links appear as no response: the other server didn’t answer the crawler. That isn’t proof the link is broken, which is why they’re separated out. You can check those by hand in the app and record the verdict, which makes the next audit more accurate.

The homepage slides

Four slides about your homepage in particular.

Homepage: access, meta, tech, linking — the standard checks on that one page: can it be reached, is its metadata in place, is its markup and language handling right, and is it properly linked.

Homepage social card — your Open Graph markup with a preview of how the link looks when someone shares it in a chat or on social media. A missing or broken card here is a small thing that shows up every single time anyone shares you.

Homepage Core Web Vitals — lab speed measurements with their thresholds: how long the main content takes to appear, how long the page is blocked before it responds, and how much the layout shifts while loading. The summary of these is the PERF score.

Homepage: rendering, extractability, E-E-A-T, schema — the deep audit. This is the one that answers “what can AI actually read here”: whether the content is present once the page renders, whether it can be extracted and quoted, whether authorship and freshness signals are there, and whether the structured data matches what the page actually is.

What to do with this block

Every failing check on these slides exists as an item in your Technical queue in the app, with the steps and — where the fix is code — a snippet ready to paste. The report tells you what’s wrong; the app tells you how to fix it. See Technical.

Order of work: blockers first, then whatever the queue ranks as high priority. Nothing else on this block matters while a blocker is failing.