Audit
LOOP audit and checklist
40 checkable items grouped by the four LOOP dimensions. Each item carries a weight. Sum your points to get a score out of 100 and prioritise the work.
Dimension L
Legibility (10 points)
Is your content extractable without friction by an AI engine?
- 01
Each page has a unique H1 carrying its primary query.
3 pts - 02
H2s segment distinct intents, not decorative sections.
3 pts - 03
Paragraphs run 3 to 6 lines, one idea per paragraph.
2 pts - 04
Sentences are standalone, readable out of context.
3 pts - 05
Every acronym is defined at its first local occurrence.
2 pts - 06
Content contains at least one table or bounded list per pillar page.
2 pts - 07
Publication and update dates are visible and authentic.
3 pts - 08
Figures are dated and sourced.
3 pts - 09
WCAG AA accessibility is respected (contrast, focus, alt, landmarks).
2 pts - 10
Critical images carry a descriptive alt (not "image.png").
1 pt
Dimension O
Ontology (10 points)
Is your brand identifiable, disambiguated and recognised as an authority?
- 01
The brand entity is named consistently across the whole site.
2 pts - 02
A schema Organization with sameAs to LinkedIn, Wikipedia, Wikidata is in place.
3 pts - 03
A factual About page establishes identity, dates and activities.
2 pts - 04
The site covers a topic cluster (1 pillar + 5 to 10 satellites).
3 pts - 05
Pages reference domain markers (competitors, tools, methods) with nuance.
2 pts - 06
Internal linking ties pillar and satellites through varied semantic anchors.
3 pts - 07
The site accrues backlinks from topically aligned sources.
3 pts - 08
External mentions (podcasts, press, LinkedIn) converge on the same identity.
2 pts - 09
A contextual (not generic) FAQ exists on pillar pages.
2 pts - 10
A glossary or definitions page exists for the domain’s key terms.
2 pts
Dimension O
Operations (10 points)
Can AI bots crawl and technically understand your site?
- 01
Explicit robots.txt for GPTBot, PerplexityBot, ClaudeBot, Google-Extended.
3 pts - 02
WAF / Cloudflare rules verified: AI bots not blocked by default.
3 pts - 03
Critical content is present in HTML without JS execution (SSR/SSG).
3 pts - 04
An XML sitemap is generated and submitted to GSC.
2 pts - 05
Every page carries a coherent absolute canonical.
2 pts - 06
schema.org structured data is validated (Rich Results Test).
2 pts - 07
Core Web Vitals: LCP < 2.5 s, CLS < 0.1, INP < 200 ms.
3 pts - 08
HTTP/2 or HTTP/3, gzip/brotli compression, HTTPS.
2 pts - 09
Images in AVIF/WebP, loading="lazy" below the fold.
1 pt - 10
llms.txt and llms-full.txt published and consistent with actual site structure.
1 pt
Dimension P
Performance (10 points)
Are you steering the effort with metrics and a cadence?
- 01
Multi-engine baseline on 10 to 20 strategic queries.
3 pts - 02
Google Search Console connected; AI Overviews report reviewed.
2 pts - 03
AI citation monitoring tool configured (Profound, Peec, Otterly, Scrunch) or manual bi-monthly monitoring.
3 pts - 04
Server logs analysed for at least GPTBot, PerplexityBot, ClaudeBot.
2 pts - 05
Monthly citation review compared against the baseline.
2 pts - 06
Quarterly pillar-page review with genuinely updated dates.
3 pts - 07
Internal log of behavioural changes observed.
2 pts - 08
Target KPIs defined (positions, impressions, share of citations).
2 pts - 09
Simple reporting shared with leadership or the team.
2 pts - 10
Documented editorial cadence: publishing frequency, refresh frequency.
2 pts
Next
You have your score. Now what?
- 1. Run the 6-step method.
- 2. Use the LOOP framework to prioritise.
- 3. Check the use cases page to adapt to your sector.
- 4. Re-test in 4 to 6 weeks, compare, iterate.