Proof: we're auditing our own sites first

In short:

  • We have no client results yet. So before auditing anyone else's business, we're running the same checks on websites we run ourselves: our own test sites.
  • So far that covers website checks. The AI-answer check on our own sites hasn't been run yet, so this page shows no before-and-after for AI answers.
  • When we run the AI-answer check, and again about 30 days after our fixes are live, we'll publish both results here, whether they improve or not.

Our test site (a site we run)

Baseline (17–18 Oct 2026)

Check Baseline What it is
Cited for its target questions 0 of 15 A search-tool proxy, not AI answers (17 Oct 2026)
Author and About page Posts credited to an author with no bio; at the time of testing, the About page described features the site didn't have weCDEF site check, 17–18 Oct 2026
Freshness 100 of the latest 100 posts showed the same "last modified" date, 17 Oct 2026, from a bulk update Site data, 18 Oct 2026

Where it stands (25 Oct 2026)

  • 88 of the latest 100 posts still show 17 Oct 2026 as their "last modified" date; 12 show dates from 18 to 24 Oct.
  • At the time of testing (25 Oct 2026), the About page still described features the site didn't have.

Source: site data and About page, checked 25 Oct 2026, 09:41 SGT.

Fixes planned and in progress

  • Rewrite the About page so it only claims what the site really does.
  • Credit posts to a real author with a short, true bio, and add an editorial policy page.
  • Stop site-wide updates from changing every post's "last modified" date, and show a visible date only when the content changes.
  • State the source and date for every figure the site publishes.

None of these is live yet, so there's nothing to compare.


The re-run we've planned

Item Plan
AI-answer baseline Not yet run. 16 questions for our test site, all naming Singapore. ChatGPT and Google AI Overviews: 3 runs each, signed out, from a Singapore connection
Re-run The same questions, same settings, about 30 days after the fixes are live. We'll post the date here once it's set
What we'll publish Both results, whatever they show: "named in X of 3 runs" for each question, share of voice, and the sources cited most

Source: weCDEF own-site prompt set and audit method.


What a client report looks like: a fictional sample

Our sample report for a fictional business shows the full layout of an audit, with made-up data for a waterproofing firm that doesn't exist. Every figure in it is fictional.

  • Readiness score: 47 out of 100 (Patchy) (fictional sample)
  • Outcome on ChatGPT: named in 4 of 12 runs, share of voice 20% of 20 firm mentions (fictional sample)
  • Outcome on Google AI Overviews: an AI Overview shown in 8 of 12 searches, named in 1 of those 8 (fictional sample)
  • Set-up: 4 waterproofing questions × 3 runs each on ChatGPT and Google AI Overviews, from a Singapore connection (fictional sample)

What we won't claim

  • We won't show any before-and-after for AI answers until the AI-answer baseline and the re-run both exist.
  • We won't present indexing changes as AI visibility.
  • We won't say our fixes caused a change when other things changed at the same time.
  • We won't use the sample report's numbers as results. They're made up to show the layout.

FAQ

Why use your own sites as proof? We don't have client results yet. Testing on our own sites first means we find the problems with our method before a client does.

Has your audit improved AI visibility for anyone? Not yet measured. The AI-answer check on our own sites is still to be run, and the re-run will follow about 30 days after the fixes are live. We'll publish both, whatever they show.

Is the sample report real? No. It's a fictional business, and it's labelled as fictional throughout.

What did the first audit find on your own sites? On our test site: an author with no bio, an About page that, at the time of testing, described features the site didn't have, and every post showing the same update date. The tables above have the details.

When will the re-run be published? About 30 days after the fixes are live. We'll post the date on this page once it's set.


Related: how we run each check · ChatGPT and Perplexity on waterproofing in Singapore · ChatGPT and Perplexity on renovation and interior design in Singapore · weCDEF home