# Alfreya Visual Similar Products

> Visual Similar Products puts a “more like this” row on the product page. AI reads each photo in your catalog for shape, colour, pattern and cut, then picks the closest pieces you have in stock, in the same gender and category.

## Key facts

| | |
|---|---|
| Status | Live with Vitruta |
| Product family | Discovery & Search |
| Where it runs | A row on the product page |
| Platform | Shopify app, or any other platform by integration |
| Languages | Turkish and English catalogs |
| Setup | Three steps: install, let AI read the catalog, add one block |
| Built for | Vitruta |
| Industries | Fashion & Apparel, Footwear & Accessories, Fine Jewelry, Watches |

## The problem

Standard “related products” lean on tags, collections and vendors, which are rarely filled in the same way twice. Someone looking at a cream knit gets a random top from the same brand. Or a men’s piece on a women’s page, which is where they stop browsing.

## What we built

AI reads the photos and details of each product for shape, colour, pattern and cut, labels its gender and category, and finds the closest pieces in the same store. Gender and category work as hard filters, sold-out pieces can be hidden, and a last AI pass puts the most convincing match first. From then on it keeps up with the catalog on its own.

## What changes

A shopper who likes a piece but not its colour, or finds their size sold out, sees a near match without leaving the page. And a women’s page shows women’s pieces.

## How it works

1. **AI reads the catalog**: Each photo is read for shape, colour, pattern and cut. A full catalog takes minutes.
2. **It labels each piece**: Gender, category and pattern, in Turkish or English, labelled automatically.
3. **It finds the closest**: The nearest lookalikes in the same gender and category, in stock, best match first.
4. **Keeps up**: New pieces, new prices and sell-outs reach the row as they happen.

## What’s included

- **Matched by appearance**: Silhouette, colour, texture and detail decide the match. Shared keywords don’t.
- **Hard gender filter**: A women’s product page never suggests menswear, and the other way round.
- **Same shelf first**: Suggestions come from the same category. They widen to the parent category only when the shelf is thin.
- **A second AI pass**: A second AI pass reorders the candidates, putting the most convincing alternative first.
- **Always up to date**: New, edited and deleted products are picked up automatically.
- **Doesn’t hold up the page**: Results come from an index prepared in advance. The row loads without delaying anything else.

## Frequently asked questions

### How is this different from Shopify’s related products?

Shopify’s suggestions lean on tags, collections and purchase patterns. This compares the photos, so a suggestion looks like the piece on screen.

### How quickly does the row appear?

Straight away. The matches are worked out in advance, and the row loads with the page without holding anything up.

## Pairs with

- [Outfit Recommendations](https://www.alfreya.com/md/en/solutions/complete-the-look): Two colours, six outfits. AI styled each one from your stock. (Live with Vitruta)
- [AI Shopping Assistant](https://www.alfreya.com/md/en/solutions/ai-shopping-assistant): Search by sentence, by photo, or by “same but in black”. (Live with Vitruta)
- [Smart Search](https://www.alfreya.com/md/en/solutions/smart-search): A search box that forgives typos, plurals and a second language. (Live with Vitruta)

---
Alfreya (Konsiyer Teknoloji A.S.) · Founded 2025 · Istanbul · info@alfreya.com
HTML: https://www.alfreya.com/en/solutions/similar-products
AI index (llms.txt): https://www.alfreya.com/llms.txt