Back to The Lab

How LLMs Choose What Products to Surface (And How Not to Be Invisible)

Photo of Bearjam
Bearjam
Guest Post

If an AI assistant (ChatGPT, Gemini, whatever) gets asked “What’s the best jacket for winter hiking?”, would your product even get mentioned? 

Showing up in AI answers is becoming as important as ranking in Google these days. The things you optimised forever? Important. But there are extra signals now – signals that decide whether you’re in the mix … or buried.

Here’s what we’re seeing, what likely matters, and what you can do today so your products aren’t ghosted.

What We Do Know: How LLMs Pick Products Now

Structured, machine-readable signals win. Products with schema (price, reviews, availability), clean categories, clear attributes, etc., are easier for LLMs to parse.

Freshness counts. If your product info is stale (old copy, outdated stock, old offers), chances are it’s less likely to be surfaced. Newer content often gets priority.

Authority & trust still matter. Domains with good backlink profiles, genuine reviews, real testimonials are more likely to be considered. Brand perception plays a role even in “AI world.”

Clarity & specificity help. Short, unambiguous product descriptions; common questions answered; bullet-lists of benefits; alt text on images – clean formatting helps.

What Probably Matters (But Not Fully Proven Yet)

Because this is still early, some of these are speculative – but they’re worth acting on:

Review volume + quality (not just star rating) might be a strong signal. The more customers trust the product, the more likely AI will pick it.

Engagement metrics: clicks, dwell time, maybe social mentions – things that show “people like this product.”

Visibility outside your site: third-party reviews, forums, brand mentions–elsewhere–could help.

What You Can Do Right Now

Because some of the powerful signals are things you can’t control easily, focus on what you can:

Make sure every product page has proper Product schema: price, availability, reviews.

Keep your product info up to date. Offers, stock, specs – don’t let them drift.

Write descriptions that answer questions customers ask – what makes this product better, what problems does it solve, what are its limits.

Use bullet points, short paragraphs, image alt text – format it so machines (and humans) can quickly “see” the value.

✓ Encourage reviews + testimonials. Even a few good, honest ones can shift things.