Cassian™ Docs
Features

LLM Readiness

How Cassian scores your store's readiness for AI-powered shopping — Layer 5 of the Cassian Score™.

LLM Readiness

LLM Readiness is the fifth layer of your Cassian Score™, weighted at 10%. It measures how well your store is positioned for the next wave of e-commerce discovery: AI-powered shopping assistants.

The way customers find and evaluate products is changing. AI systems embedded in search engines, browsers, and voice assistants are increasingly answering shopping queries — "What's a good running shoe for flat feet under $150?" — and returning specific product recommendations rather than a list of links. Stores that are structured for AI readability appear in these results. Stores that aren't, don't.

LLM Readiness is the lowest-weighted layer, but it's often where the fastest wins are. Many of the fixes are one-time tasks that take less than an hour.


What Cassian checks

All LLM Readiness checks are deterministic — they run instantly on every scan and produce consistent results. These issues appear under the LLM Readiness filter on the Issues page.

Product schema

Structured data (schema markup) is the primary way AI systems extract machine-readable information from web pages. On every crawled product page, Cassian checks for Product JSON-LD schema:

  • Missing Product schema (High severity) — the product page has no Product schema at all. AI shopping assistants and search engines rely on structured data to understand products.
  • Incomplete Product schema (Medium severity) — Product schema exists but is missing one or more of: name, description, image, offers/price, brand, or a product identifier (SKU, GTIN, or MPN).

Partial schema is better than no schema, but each missing attribute reduces the confidence AI systems have when representing your products.

Where to fix: Most Shopify themes include Product schema by default. Check by viewing page source and searching for "@type": "Product". If it's missing or incomplete, add it via a theme customisation or a structured data app.

Organisation schema on your homepage

Cassian checks your homepage for Organization, WebSite, Store, OnlineStore, or LocalBusiness JSON-LD schema. A homepage with none of these is flagged as High severity — this schema is how search engines and AI assistants identify your business.

This is a presence check: Cassian confirms the schema exists but does not currently validate the completeness of its fields.

Pages without BreadcrumbList schema are flagged at Low severity. Breadcrumb schema helps search engines and AI systems understand your site hierarchy — which collection a product belongs to, and how your catalogue is organised.

Robots meta tag

Pages without a <meta name="robots"> tag are flagged at Low severity. The tag isn't required, but it helps search engines and AI crawlers understand your indexing preferences.

JavaScript-heavy pages

Content that only renders after JavaScript execution may not be visible to AI crawlers that don't execute JavaScript — common with headless setups or apps that load content asynchronously. This isn't a scored check: when Cassian detects that a page needs JavaScript to render its content, it re-crawls that page with rendering enabled so the analysis still sees your full content.

llms.txt — generated for you, not scored

llms.txt is an emerging standard — a markdown text file placed at the root of your store (e.g., yourstore.com/llms.txt) that tells AI systems what your store sells, who you are, and how to interpret your catalogue. Think of it like robots.txt, but for AI assistants rather than search engine crawlers.

Cassian does not score your store on whether an llms.txt file is present. Instead, the AI Product Feeds feature (Analyst plan and above) generates a complete llms.txt for you — including your store summary, category counts, and up to 50 featured products — ready to download and deploy to your own domain root.

Answer-ready formatting

When an AI assistant answers a shopping query, it draws from the text it can extract from product pages. Content written in complete, natural language sentences is significantly more usable than the same information presented as bullet-point attribute lists.

Cassian AI™ flags pages where product descriptions consist primarily of specification bullets without any prose content. These findings are part of the Content Quality analysis, so they appear under the Content filter on the Issues page — but they matter for AI readiness too.

Example: low answer-readiness

- Material: 100% cotton
- Weight: 280gsm
- Care: Machine wash cold
- Origin: Portugal

Example: high answer-readiness

Woven from 280gsm 100% Portuguese cotton, this t-shirt is substantial enough to hold its
shape after repeated washes. Machine washable cold, it softens with wear rather than
pilling. If you're after something that lasts years rather than seasons, this is it.

Both versions contain the same information. The second version is far more usable by an AI assistant answering "What are the best quality basics for men?" The spec-only version is not usable at all.


How important is LLM Readiness?

LLM Readiness is weighted at 10% — it's the lowest-weighted layer. A poor LLM Readiness score won't tank your overall Cassian Score. But the trend is clear: AI-powered shopping is growing, and the stores that are positioned for it now will benefit compoundingly as adoption increases.

The good news: most LLM Readiness improvements are one-time tasks. Adding Organisation schema, completing your Product schema, and ensuring your descriptions have prose content — once done, these improvements persist.

Quick wins (under 1 hour each)

Add Organisation schema to your homepage. Deploy the llms.txt generated by AI Product Feeds. Write intro paragraphs for your top 10 products.

Medium effort (1–4 hours)

Add Product schema to all product pages (if missing). Complete missing Product schema fields (brand, SKU, image). Add breadcrumb schema.

Ongoing

Write new product descriptions with prose content from the start. Redeploy your llms.txt as your catalogue evolves.


Where to see LLM Readiness results

  • Issues page → filter by "LLM Readiness" category. All LLM readiness issues listed by severity.
  • Scans page → click any scan → LLM Readiness section shows your layer score and all issues found.
  • Dashboard overview → the LLM Readiness card shows your current layer score.

Common questions

How do I add an llms.txt file to Shopify? Start with the AI Product Feeds feature — Cassian generates a complete llms.txt from your catalogue, ready to download. To deploy it in Shopify, the most common approach is to create a new page with the content, then configure a custom URL redirect from /llms.txt to that page. Alternatively, some themes support adding custom files directly, or you can use a Shopify Files app.

Will fixing LLM Readiness improve my search engine ranking? Not directly — traditional search engine rankings and AI-powered search surfaces are different systems. Improving LLM Readiness affects your visibility in AI-generated answers (in search engines that use AI overviews, browser-integrated assistants, voice search, etc.), not your traditional organic ranking. That said, several LLM Readiness improvements — Product schema, breadcrumb schema, Organisation schema — also benefit traditional SEO.

Is llms.txt actually a real standard, or is Cassian inventing it? It's a real and growing standard, proposed by prominent developers in the AI and web communities. Major AI systems have begun supporting it. It's in the same early-adoption phase that robots.txt was in the 1990s — not yet universal, but moving in that direction. Stores that adopt it early have a head start.

My LLM Readiness score is 45 — should I drop everything to fix it? At 10% weight, a 45% LLM Readiness score reduces your Cassian Score by about 5.5 points compared to a perfect layer. Focus on Technical Health and Content Quality first if those layers are also below target. Then come back to LLM Readiness for the quick wins.

I have Product schema on my product pages but Cassian still flags them — why? The flag is likely "Incomplete Product schema" rather than missing schema. Cassian checks for name, description, image, offers/price, brand, and a product identifier (SKU, GTIN, or MPN) — the issue detail lists exactly which fields are missing. Malformed JSON-LD (a common issue with manually written schema) can also prevent the schema from being read at all. Use Google's Rich Results Test tool to validate it.

My store is JavaScript-heavy — does that hurt my LLM Readiness score? Not with Cassian: when a page needs JavaScript to render its content, Cassian re-crawls it with rendering enabled, so the analysis sees your full content. But it's worth knowing that some external AI crawlers don't execute JavaScript — content that only appears after rendering may be invisible to them, regardless of your Cassian Score. If your Shopify store uses a standard theme, this is usually not an issue.

On this page