Reviews AI agents can actually read: what changes when your stars are in the HTML

The reviews platform that is natively readable by AI agents, and that defaults to free, has a boring mechanical trick: the stars are in the HTML before any script runs.

That is the whole product claim. Not that ChatGPT will pick you. Not that Google will rank you. Can AI assistants see my Shopify product reviews? already documented the gap. This post is what changes when the library itself is built for that gap.

What a fetcher gets today

Open a product with a typical review widget, view source, and search for a sentence you know a customer wrote. Inspect Element will show it. View Source often will not.

The widget was built for shoppers. The page HTML loads, JavaScript calls the review app, and quotes appear. Humans wait a beat. Assistants that fetch once leave with a shell.

Review data vs review widgets splits the jobs. Display is one job. Machine-readable ratings are another. Most apps optimize the first and treat the second as a rich-snippet afterthought.

What AgentReady puts in the first paint

Approved reviews are snapshotted into Shopify metafields. The theme blocks render that snapshot in Liquid. No AgentReady request, no Vercel request, no database round trip is required to show the initial set.

Show more reviews is a later page, loaded through a signed endpoint on your domain. Continuation can use JavaScript. First paint cannot depend on it. The how-to is AI readability.

JSON-LD is a second surface

AggregateRating is emitted only when no other app already owns review markup. If a competitor's snippet is present, AgentReady demotes its own. One rating story per page. Details in One AggregateRating per page and the ownership doc.

On paid plans the same approved ratings also flow into agents.md, llms.txt, and the catalog feed. Those files help an agent find the number. They are not a ranking lever.

Free by default

Storefront display, import from Judge.me, Okendo, Loox, or CSV, moderation, public replies, photo re-hosting, and export stay on the free tier. That is what free means here.

If you are mid-migration, start with the migration checklist, not with uninstalling the old app.

The sell page is AgentReady Reviews.

Frequently asked questions

Can AI agents read Shopify product reviews?

Only if the text is in the first HTML response or in structured data. Widgets that fetch reviews after paint leave a fetcher with an empty shell. AgentReady Reviews renders approved reviews from Shopify metafields with zero JavaScript on first paint.

Does making reviews readable to AI improve ranking?

No. Readable ratings help assistants see and read your ratings. They do not improve ranking, citations, traffic, or revenue, and they do not mean an assistant will name your store.

Is AgentReady Reviews free?

Storefront display, import, moderation, replies, photos, and CSV export are free. Ratings in agents.md, llms.txt, and the catalog feed are on paid plans.

What does AgentReady Reviews not replace?

Judge.me, Okendo, and Loox still lead on collection emails, video, question boards, and photo carousels. AgentReady now ships a storefront write form: new reviews arrive unverified and pending. Request emails stay a supervised pilot.

Get found and recommended by AI shopping assistants.

Run the free AI-Readiness Checker to see, in about ten seconds, how ChatGPT, Perplexity, and Google read your store today and exactly what is holding it back. Then AgentReady fixes the gaps for you, adding Schema.org structured data, an llms.txt directory, and an ongoing audit. Install free; every software feature is $29/mo.