Book a demo

Marketing

Shoppers search for it. Your product data doesn't say it.

Cernel fills the attributes shoppers search for and writes your product copy in your brand voice, for every market you sell in.

A sculpted cream ceramic table lamp with a linen shade on an oak sideboard, beside a stack of books

ceramic table lamp linen shade

yourshop.co.uk › lighting › table-lamps

ARC LAMP CER 44CM WHT

Lamp. Base: ceramic, white. Shade: fabric. H 44 cm. E27.

Your supplier's text: none of the shopper's words

Trusted by e‑commerce teams at

Sound familiar?

The campaign is ready. The product data isn't.

Products don't show up for what shoppers search.

  • Feeds get rejected for one missing field.
  • Copy starts from zero, product by product.

Cernel fills what shoppers search for and writes copy in your voice, in every market.

39%
of ecommerce sites have issues with feature searches, like a product by its material or propertyBaymard Institute, search query types, 2026
38%
of shoppers will abandon a site if they can't easily find key features like sizing and materialsAkeneo survey of 1,000 US shoppers, Sept 2025
22%
of shoppers use AI tools to research products before they buySalsify Consumer Research, 2026

Found for the shopper's word

Your supplier says brick. Shoppers search for rust.

Products go live in the supplier's words, and the shopper who searches for rust lands in another shop that says rust. Cernel reads what shoppers in each market search for and shows you the word that wins. You add it to your colour list once; from then on an agent sets it on every product it fits, in every language you sell in.

Two linen cushions, one white and one rust, stacked on a round wooden stool against a white wall

From the supplier

CUSHION COVER LIN 50X50 COL. BRICK

Monthly searches

  • brick cushion cover70
  • rust cushion cover2,900
ColourRust From your value list
  • UKRust
  • DERost
  • DKRust
  • SERost
  • FRRouille
  1. Reads your supplier's wordThe file says brick, so the shop says brick.
  2. Checks what shoppers searchMonthly searches per market: rust wins by far.
  3. Sets the shopper's wordRust goes on your colour list, and the agent fills it in.
  4. In every marketThe same colour in each language you sell in.

Example product and search figures.

Google Shopping

Rejected in Merchant Center, or approved and running on half a feed.

A missing brand or a wrong category, and the product drops out. The rest go live with a thin title and no colour or material, so Google has to guess, and your ad budget pays for listings that can't compete. Cernel fills every field the feed asks for, product by product, puts each product in Google's category, and exports the catalogue as one feed file, ready to upload to Merchant Center.

CernelSuede trainers, gum sole, Google Shopping feed itemFeed file ready
Feed fieldValueFrom
id, gtin, brand, price, sizeFrom your shop or supplier fileYour data
google_product_categoryApparel & Accessories > ShoesYour category
product_typeShoes > Trainers > Suede trainersAgent
titleSuede trainers with gum sole, off-white, unisexAgent
colorOff-whiteAgent
materialSuedeAgent
genderunisexAgent
age_groupadultAgent
product_highlightSuede upper on a gum rubber soleAgent
custom_label_0AW26 launchYour label
Exported with the catalogue as one Google Shopping feed file, ready to upload to Merchant Center.

Example product and values. Field names are Google's product data specification.

Your voice, every market

Change a rule once. Every description follows.

Your tone of voice, brand rules and the words you never use live in one shared library, and every agent reads them when it writes. You try it on your own products before it touches the catalogue. Per language you choose to write fresh or translate, each with its own instructions, so copy reads like it was written for that market.

Data Sources / brand / voice

  • Warm and plain, in short sentences
  • Lead with the material
  • Words we never use: perfect, stylish

Read by every agent that writes for you, in every language

  • A sculpted cream ceramic table lamp with a linen shade on an oak sideboard, beside a stack of books

    Arc table lamp

    Before: The perfect lamp for a sideboard: a sculpted ceramic base under a natural linen shade.

    After: A sculpted ceramic base under a natural linen shade, for a soft light on a sideboard.

  • A pair of off-white suede trainers with gum soles, one resting on a green shoebox, on a pale blue ground

    Suede trainer

    Before: Stylish off-white suede on a gum sole, perfect for every day.

    After: Off-white suede on a gum sole, made to wear every day.

  • Two linen cushions, one white and one rust, stacked on a round wooden stool against a white wall

    Linen cushion cover

    Before: A stylish cushion cover in washed linen, 50 x 50 cm, with a hidden zip.

    After: Washed linen in a deep rust, 50 x 50 cm, with a hidden zip.

Example rule, products and texts.

AI search

Shoppers ask an assistant now. It can only compare what your data says.

A shopper asks ChatGPT or Google AI Mode in a full sentence, and the assistant compares products on facts. No one decides what an assistant recommends, but rich, structured product data helps it find and compare yours. That is what Cernel fills: the roast, the brew method and the weight, each its own attribute, not hidden in a paragraph or missing.

A shopper asks an AI assistant

Whole bean coffee for my filter, a light roast, something fruity. A 250 g bag.
A plain kraft paper coffee pouch on a white surface in soft daylight

Ethiopia Sidamo, whole bean

Form
Whole bean
Brew method
Filter, Pour-over
Roast
Light
Tasting notes
Peach, Bergamot
Weight
250 g

Every fact in the question is in the product data

Example question, product and values.

What changes

Your team sets the rules. Cernel does the product-by-product work.

Where your content goes

  • Your shop and your PIM: descriptions, meta titles, values and translations, sent back once approved
  • Google Shopping: a feed file, ready to upload to Merchant Center
  • Anything else: a file export, or the API

What you control

  • Brand voice, rules and keywords in one shared library, read by every agent
  • A test run on your own products before anything touches the catalogue
  • You choose which values go live automatically and which get a review step; review steps check every value and only ask a person when something fails or needs a decision

What you see

  • Search demand: how much of what shoppers search for your data answers, per market
  • The options and attributes worth adding, ranked by missed searches
  • Every AI-written value with its inputs, reasoning and prompt

What teams get back with Cernel

All customer stories

  • Molly & My

    Molly & My builds every market on one foundation

    products enriched across 8 markets
    20,000
    faster time to market
    4x
    organic traffic to product pages
    +29.3%
    Read the story
    A stack of folded knitwear in low warm sunlight

  • Firtal Group

    Firtal Group added a thousand extra hands without adding headcount

    products enriched
    +25K
    organic traffic to product pages
    +20%
    faster time to market
    +80%
    Read the story
    In the warehouse at Firtal Group

  • Tiffany Lighting Direct

    Tiffany Lighting Direct gets a new supplier range live in days, not months

    to get a new supplier range live
    Days
    products from 20+ brands
    13,000+
    extra hires needed
    0
    Read the story
    A stained-glass table lamp on a wooden table in a quiet cafe

FAQ

Questions marketing teams ask us.

  • Will it sound like our brand?

    Every agent reads your tone of voice, brand rules and terminology from one shared library, and you can add standing instructions for each kind of text. You try an agent on your own products before it touches the catalogue. Per language you choose to write fresh or translate from another language, each with its own instructions.

  • Where does the search data come from?

    Search demand uses monthly search figures from Google Ads, read per market (every language you add is a market), and matches each search term to the attributes and values shoppers use. Brand searches are left out. Its recommendations show which options or attributes to add first; you review and save them.

  • Can Cernel write our meta titles and descriptions too?

    Yes. Agents write product descriptions, meta titles and meta descriptions from each product's own facts, to the length, structure and keywords you set, in every language you sell in. Search demand shows which words shoppers use in each market, so you can put them in the rules. Once approved, the copy goes back to your shop or PIM, or out as a file.

  • Will this get us recommended by ChatGPT or Google AI Mode?

    No one can control what an assistant recommends. Rich, structured product data is proven to help, because it gives the assistant the facts it reads and compares. That is exactly what Cernel provides: every fact shoppers ask about filled as its own attribute, and descriptions written from those facts, on every product page.

  • Does a person have to review every value, and can we stop claims we can't make?

    You choose which values go live automatically and which get a review step; review steps check every value and only ask a person when something fails or needs a decision. Add a review step with a rule like 'fail if it calls the product sustainable without a certificate', and a failing value is held for a person or rejected. A person sees every held value with its inputs, reasoning and prompt.

Where to next

See how each fix works.

Bring your brand guide and ten products.

In a demo we run your brand voice on your own products, in the languages you sell in, and show you what shoppers in each market search for.

Book a demo