Real Conversations With AI

Gemini's AEO optimization suggestions

Date July 24, 2026 Engine Gemini
Four findings on how well ZenMasterWorks executes on its own AEO principles. Three checked out as accurate or reasonable. The second one — described as "mixed execution" — turned out to be a much bigger gap than the wording implied. Logged on the incident log as entry 11.156.
Gemini

An evaluation of the site against those optimization strategies reveals a strong alignment with some practices, while highlighting clear opportunities to tighten others:

1. Machine-Readable Indexing via llms.txt

Current Status: Incomplete / Inaccessible in typical retrieval paths. While having a dedicated machine-readable overview is an ideal vector for AI answer engines, ensuring it is cleanly exposed and explicitly referenced in your robots or sitemap structure makes it more reliably consumable by automated crawlers.

Confirmed real: llms.txt existed but was referenced nowhere — not in robots.txt, not in sitemap.xml, no crawlable link anywhere on the homepage. Fixed with a real link in the homepage head and footer.

2. Granular Schema Markup

Current Status: Mixed execution across pages. While technical blueprint and methodology pages make strong use of structured data (like ItemList and BreadcrumbList), checking specific child pages to ensure they utilize precise types (such as TechArticle or BlogPosting rather than generic wrappers) prevents parsing ambiguities.

Confirmed real, and larger than described: 17 blog posts had zero Article or BlogPosting schema at all, not just imprecise typing — including two posts built earlier that same day. Fixed across all 17.

3. Direct Cross-Referencing & Semantic Anchors

Current Status: Implemented effectively in localized flows. Internal paths connect high-value pages together, but ensuring high-intent landing pages link directly back to accountability assets like the incident log helps build a clear semantic trail for retrieval models.

Confirmed real: three landing pages (blueprints, technical foundation, resume-to-website) had zero links to the incident log. Fixed on all three.

4. High-Fidelity Content Structure

Current Status: Strongly executed. The site utilizes clean HTML landmarks (<main>, explicit heading hierarchies) and clear, descriptive copy directly beneath headings, making it straightforward for modern RAG architectures to segment, extract, and attribute specific paragraphs without parsing noise.

General praise, not independently re-verified line by line — no specific claim to fact-check here.
← Back to all conversations