ChatGPT, Gemini, Claude, Perplexity and Grok can't recommend what they can't see

Sighted scores every product for AI assistants, fills its gaps from your data and checks mentions weekly.

Get your free scan

Free, no account, a report in a few minutes.

Works with

ShopifyGoogleChatGPTGeminiperplexity

Shoppers ask AI what to buy. It recommends the few products whose data answers them.

If your blazer does not say linen, sand and medium, it is not in the answer.

I need a smart linen blazer for the office, warm sand, size medium, under 150 pounds, something I can wear to dinner too

The Relaxed Linen Blazer in Sand fits your brief. It is 100% linen with a relaxed fit made to go over a shirt or a tee, so it works at the office and at dinner. It costs £129.yourstore.com

Example product: a relaxed sand-coloured linen blazer worn over a white tee

Relaxed Linen Blazer in Sand

£129 · Medium in stock

higher conversion from shoppers sent by AI assistants
60%
Adobe Digital Insights, July 2026
of shoppers use AI while buying
45%
IBM Institute for Business Value, January 2026
more orders from AI referrals, year on year
3x
Shopify, August 2026

AI assistants skip products they cannot read. Here is how Sighted fixes that.

Scored, filled from your own data, approved by you, published, measured weekly.

Example: the catalogue ranked by impact, with its readiness, 63 of 100.

Score every product

Scan

Each product gets a readiness score, a grade and its gaps, ranked by impact.

Example: the same product before and after, 41 (D) to 87 (B), its gaps filled from the store's own data.

Fill the gaps from your data

Enrich

Material, fit, colour and shopper FAQs, written from your descriptions, options and photos, in your voice.

Example: three rewrites waiting in the review queue, each with Approve and Restore original.

Approve before it goes live

Review

New titles and descriptions wait in a review queue. One tap approves, one restores your original.

Example: every channel ticked, beside the product's Merchant Center feed row.

Publish where assistants read

Publish

Your store, Google Merchant Center and the ChatGPT and Perplexity feeds, from one approved catalogue.

Example: a rewrite's claims with their sources; the unsourced one is struck through and held.

Never invents a fact

Verify

Every claim is traced to your data. Anything Sighted cannot source is held for your review.

Example: the Results page over 8 weeks, from GA4, Search Console, Shopify and Sighted's checks: readiness 41 to 87, AI assistant sessions 12 to 141 a week, 4.8k search impressions a week, 25 of 69 answers naming the store.

Prove it is working

Monitor

Readiness, sessions from Google Analytics, searches from Search Console, and how often ChatGPT, Perplexity and Gemini name you.

Results from retailers using Sighted.

sales
+36%
Onboard Sportswear
organic conversion
+64%
The Game Collection
traffic
+200%
The Bean Coffee Company
AI citation rate, up from 0%
27%
ChristianShirts

Trusted by retailers and SO-ME partners.

  • Marks & Spencer
  • River Island
  • Onboard Sportswear
  • New Look
  • French Connection
  • Juicy Couture
  • The Bean Coffee Company
  • Debenhams
  • Kaiia the Label
  • Lavish Alice
  • EGO
  • Blue Vanilla
  • Public Desire
  • Lemon Lunar
  • Gini London
  • OMNES
  • Never Fully Dressed
  • Yours Clothing

Get discovered. Get ready to be bought.

Sighted is working with PayPal so shoppers can buy your products inside AI assistants, not just find them. Enrich your catalogue today and it is ready when checkout arrives.

Ask us about agentic checkout
PayPal

Questions

What merchants ask before they leave an email.

All questions

Assistants such as ChatGPT, Perplexity and Gemini match a shopper's question against the product data they can read: your product pages, their Schema.org markup and your shopping feeds. A product whose data states the material, fit, size, price and stock can be matched to a detailed question. A short title and a thin description usually cannot.

No. Sighted scores your products without touching them, and you can preview a rewrite on one product before anything runs. New titles and descriptions wait in a review queue until you approve them, unless you turn on automatic publishing; one tap restores your original. Only approved rewrites reach your store and your feeds, and a field you edit yourself stays yours.

No. Every claim in a rewrite is checked against your own product data and photos before you see it. A claim Sighted cannot support, such as a material or a measurement, is held for your review instead of published, and FAQ answers that would guess at a typical value are dropped. New titles and descriptions wait for your approval, and a field you edit stays yours.

Every product gets a score out of 100 and a grade from A to F. The score weighs six things: structured data assistants can read, real detail in your title and description, data that agrees with itself, AI crawler access to your pages, feeds ready for the assistants' channels, and stated reviews and store policies. It updates on every change and every rescan.

A few minutes. Connect your store, or upload a CSV or a product feed URL, and scoring starts as your products come in. Nothing changes in your store until you approve a rewrite.

On a schedule (weekly on Pro, monthly on Growth), Sighted asks the assistants it tracks, such as ChatGPT, Perplexity and Gemini, a set of real shopper questions about the kind of products you sell, and records whether each answer names your store. You see the results by assistant and over time on the Results page, so you can tell whether your changes moved anything.

Invisible to AI? Get Sighted.

Early access scores your whole catalogue and previews one rewrite before anything runs.