laurelinlabs.com · 12 pages · 08 July 2026 · by Laurelin Labs, on ourselves

How much of your site would an AI bother to quote?

Search has quietly changed shape. Google, ChatGPT and Perplexity now write the answer at the top of the page, and they build it from a handful of sources they decide are worth quoting. Everyone else is scenery. If you are not one of the few, the answer about your own field gets written without you.

So we pointed the tool at our own site first. It would be a bit rich to grade your content if we had not graded ours. Here is what it found, warts and all, and what we are fixing. If a score is worth trusting, it should be willing to embarrass its makers.

Every score below quotes its evidence. Hover any dotted term for a plain explanation. The maths is all there too, inside 'More information' on each row.

The headline

41.1/ 100

We took our own medicine

Answer first: we practise what we preach on identity, and our weak spot is hard data. Trust signals score 53, our strongest area, because we wire our author and organisation entities properly, the exact work most sites skip. Uniqueness (45) and personal experience (44) are healthy: original teardowns, our own tools, first-hand builds. The honest gap is real, checkable data at 19, many of our pages explain ideas without enough numbers on the page. We are taking our own advice and fixing it in public.

If it helps you trust the score: the tool did not flatter its makers. It found our weakest pillar and named it. The searches we teach about are already answered by AI, using Reddit threads and Google's own docs. Nobody is exempt, us included.

Pillar 1 · counts most

45 /100

Content uniqueness

Healthy. Our teardowns and framework pages add genuinely new angles, though a couple of blog posts sit closer to the crowd.

Pillar 2

44 /100

Personal experience

Real. First-hand tool builds and 'we nearly shipped a bad spec' honesty read as lived work, not desk research.

Pillar 3

19 /100

Real, checkable data

Our weak spot at 19. Many pages explain concepts without enough hard figures on the page. We are adding data to our own posts.

Pillar 4 · our strong suit

53 /100

Trust signals

Our strongest area at 53, because we do this deliberately: linked author and organisation entities, sameAs to real profiles. It is the cheapest, highest-leverage fix for most sites, and the one we lead with.

Page by page

Every score, in plain English, and what to do about it

Click any row (or the button) for what each score means for you and the specific fixes. The technical evidence, embeddings, redundancy, schema detail, sits under 'More information'. Rows marked AI answer face a search where Google already shows an AI answer, so the citation slot is live now.

Page	Score	Uniqueness	Experience	Real data	Trust

Where you stand today

We win our name and our tools, not yet the big terms

We pulled the live Google UK results for all 12 searches and measured share of attention, weighted by position.

9 / 12searches already showing an AI answer, so the citation slots are live

53 / 100our trust-signal score, our strongest pillar and the one we recommend first

19 / 100our data score, the weak spot the tool caught in its own makers

On the competitive terms we teach about, the attention is held by Reddit, Google's own developer docs and Ahrefs. We win our brand and tool names; the broader terms are exactly what our own Phase 2 tackles. We are on the same journey we would put you on.

What it is worth

Proof of method

We eat our own cooking

This is a self-audit, not a sales page. If a tool is worth trusting, it should grade its makers first. Then we would happily point it at yours.

Talk to Piers about this →

45 / 44 / 19 / 53our four scores today: uniqueness, experience, data, trust

+ dataour own Phase 2: hard numbers on concept pages

dayswhat Phase 1 takes, on our site and yours

your site →the same audit and plan, next

We are running the same three-phase plan on ourselves that we would run for you. Watch this score move over the coming months, we will publish the change either way.

The opportunity, with the working shown

Why this matters, for us and for you. The same three effects apply to any site on these terms.

1 · Clusters, not keywords

Each page competes for a neighbourhood of searches

AI splits one question into many related ones (query fan-out). A page that owns its topic surfaces across 20 to 50 of them.

2 · The gap is the opening

The pages winning these terms are beatable

Reddit threads and generic explainers hold a lot of this attention. A source with first-hand material and a verifiable identity is exactly what gets picked instead.

3 · A cited page compounds

Trust is an asset, not a campaign

A page that becomes the cited answer earns trust, links and enquiries long after it ships. That is why we treat this as infrastructure, not a one-off.

This is a self-audit, so there is no client value model here. The point is simpler: we hold ourselves to the standard we sell. If you want the same audit and plan on your own catalogue, that is the next step.

The plan

Three phases, cheapest and surest first

The same three phases we would run for you, run on ourselves, in the open. Phase 1 is days.

Phase 1 · weeks 1-2 · ~3 days

Finish wiring our own identity

Add the last of our person and organisation markup on the few pages still missing it.
Confirm every author byline links back to the canonical /about entity.
Tidy sameAs so every profile we control is connected and none we do not are claimed.

Phase 2 · weeks 3-6 · ~4 days

Fix our weak spot: put hard data on concept pages

Add one real figure or dataset to each framework and blog post that currently explains without evidence.
Turn our tool results into published, checkable numbers.
Lead each page with the answer-first callout carrying a fact, not just a claim.

Phase 3 · weeks 6-10 · ~3 days

Track and re-score, publicly

Track rankings and AI citations across our clusters.
Re-run this score monthly and publish the delta.
Feed the gaps into the next quarter's content.

Roughly 10 days across 10 weeks, on our own site, in the open. The same plan is what we would run for you.

Start with Phase 1 →

Piers Butler, founder of Laurelin Labs and an AI-search consultant. We audit against Google's own primary sources and our Four Pillars framework, we prototype in the open, and we published the teardown of this very tool's first draft. That is the standard we would bring to your work.

How it was measured

No black boxes

Answer first: every number here came from your live pages and the live Google results, measured the same way for every page, with the workings on show.

The four things we measure, in one line each

Content uniqueness. We pull the live top results for the page's search and measure how much the page says that those results do not already say. New only counts if it is on-topic.

Personal experience. A judge reads the page for signs a real person did the work, and quotes the lines that prove it.

Real, checkable data. We count genuine figures and units per 100 words. Bare dates do not count. Hub and index pages aggregate many teasers, so their data score describes the listing, not one article.

Trust signals. We read the page's structured data and outbound links, and count only those that resolve to a source Google recognises. A site's own social profiles do not count.

Visibility and the value model

For each search we pulled the full live Google UK results including AI answers, weighted each site's appearances by position on a standard click-through curve, and summed them into the attention chart. Volumes and click prices come from DataForSEO.

What this does not claim

The way the four scores combine is a starting model, not yet tuned against outcomes, so the total is directional. For this run the uniqueness comparison uses the top result snippets rather than full competitor pages. AI answers come and go, so their presence is a snapshot. Rankings were captured on 08 July 2026 and move constantly.