Dog-fooding Our Own Tools

62 to 99 in One Session

We ran the AEO Toolkit on our own GitHub Pages site, followed the recommendations, and tracked every change. Here are the real results.

62
Before
99
After

Where the Points Came From

The audit scores four dimensions. Here's how each changed.

Dimension Before After Change
AI Crawler Access 12 / 25 25 / 25 +13
Structured Data 16 / 30 29 / 30 +13
Content Citability 21 / 25 25 / 25 +4
Entity & Authority 13 / 20 20 / 20 +7
Total 62 / 100 99 / 100 +37

Every Signal, Before and After

Green means the signal passed. Red means it failed. The audit checks 24 signals across all four dimensions.

Signal Before After
AI Crawler Access (12/25 → 25/25)
robots.txt existsFAILPASS
GPTBot not blockedPASSPASS
ClaudeBot not blockedPASSPASS
PerplexityBot not blockedPASSPASS
Google AI crawlers not blockedPASSPASS
llms.txt presentFAILPASS
llms-full.txt presentFAILPASS
sitemap.xml accessibleFAILPASS
Structured Data (16/30 → 29/30)
JSON-LD presentPASSPASS
@graph connected entitiesFAILPASS
Organization schemaPASSPASS
FAQ or HowTo schemaPASSPASS
Breadcrumb schemaFAILPASS
Open Graph tags completeFAILPASS
Twitter Card tagsFAILPASS
Content Citability (21/25 → 25/25)
Single H1 tagPASSPASS
300+ wordsPASSPASS
3+ heading levelsPASSPASS
Short paragraphsPASSPASS
Q&A heading patternsPASSPASS
Lists presentFAILPASS
Tables presentFAILPASS
Meta description 120-160 charsPASSPASS
Image alt text ≥ 80%PASSPASS
Entity & Authority (13/20 → 20/20)
sameAs social linksPASSPASS
Author/Person schemaFAILPASS
Contact page linkedFAILPASS
About page linkedFAILPASS
Brand consistencyFAILPASS
Why the last point? The single missing point is in Structured Data — one optional schema enrichment the audit recommends. Every actionable signal now passes. The original GitHub Pages subdirectory limitation (robots.txt, llms.txt, sitemap.xml unreachable at the domain root) was resolved by moving to a custom domain.

What We Actually Changed

Every fix came directly from the audit's recommendations. No guesswork, no SEO voodoo. Here's the full list.

Added @graph JSON-LD

Connected our Organization, Person, SoftwareApplication, and BreadcrumbList schemas into a single @graph structure instead of isolated JSON-LD blocks.

+5 structured data

Added OG and Twitter Card Meta

Added og:image, og:site_name, twitter:card, and twitter:image tags for complete social sharing metadata.

+4 structured data

Added BreadcrumbList Schema

Created a three-level breadcrumb: Boxford Partners → Tools → AEO Toolkit. Tells AI crawlers exactly where this page sits in the site hierarchy.

+4 structured data

Added Person Schema

Added a Person entity for Patrick Mitchell (Founder) linked to the Organization via worksFor. Author field on SoftwareApplication references the Person.

+3 entity authority

Added About and Contact Links

Added navigation links to the About page and Contact page on boxfordpartners.com. Entity authority checks for these as trust signals.

+6 entity authority

Fixed Brand Consistency

Aligned the Organization name in JSON-LD with the brand name extracted from the page title and og:title, so all signals return "AEO Toolkit".

+4 entity authority

Added Lists and Tables

Added a summary list of the three scoring dimensions and a table showing the scoring breakdown. AI assistants prefer structured content they can directly cite.

+4 content citability

Created robots.txt, sitemap.xml, llms.txt

Created robots.txt with AI bot policies, sitemap.xml, llms.txt, and a full llms-full.txt reference document in .well-known/. Added .nojekyll to ensure GitHub Pages serves dotfiles.

+9 crawler access

Moved to Custom Domain

Migrated from patrickboxfordpartners.github.io/rote-playoffs/ to aeo.boxfordpartners.com. This resolved the subdirectory path limitation that prevented crawlers from finding root-level files like robots.txt and sitemap.xml.

+4 crawler access

37 Points From One Session of Changes

Every change on this page was driven by specific, actionable audit findings. No SEO expertise required. No vague advice like "improve your content." Each recommendation told us exactly what to add, where to add it, and how many points it was worth.

We initially hit a ceiling at 86/100 because GitHub Pages project sites serve from a subdirectory — crawlers couldn't find robots.txt and sitemap.xml at the domain root. Moving to a custom domain (aeo.boxfordpartners.com) resolved that constraint and unlocked the remaining 13 points in AI Crawler Access, bringing us to 99/100.

This is what the AEO Toolkit does. It tells you exactly what AI assistants see when they look at your site, exactly what's missing, and exactly how to fix it. Run it on your own site and see where you stand.

Run This on Your Own Site

Takes under a minute. No account, no API keys, no data leaves your machine.

Get Started View Source