Shoppers ask ChatGPT what to wear and what to put on their skin · Sayrank gets your products named · Scan the site, fix what blocks the crawler · The agent writes, you approve · Four engines, daily runs · Shoppers ask ChatGPT what to wear and what to put on their skin · Sayrank gets your products named · Scan the site, fix what blocks the crawler · The agent writes, you approve · Four engines, daily runs ·

Sayrank

Start free trial

The whole loop: read, fix, write, prove

Get your products named when shoppers ask AI

Sayrank checks whether ChatGPT, Claude, Gemini and Perplexity can read your store at all, fixes what blocks them, rewrites the cards so there is something to quote, and tracks every answer where your products appear. The agent prepares each change and waits for your approval.

Start free trial See the product

14-day trialNo credit cardFirst scan in two minutes

Scancan engines read you
Fixyou approve, agent writes
Fillcards worth quoting
Trackwho is named instead
Provebefore and after

4.4xA visitor who arrives from an AI answer is worth 4.4 times an organic search visitor. Sayrank shows whether the engines name you at that moment.Semrush, 2025

Catalog agentSKU TA-2210 · Black resin-handle bagScanning

Color black Leather smooth calfskin Handle green marbled resin Shape soft, gathered front Trend resin top handle +47%

Card scoreEngines cannot quote this card

Scan image and attributes
Match 11 shopper prompts
Rewrite card in brand voice
Export to Shopify CSV

Who it is for

Compared with

Answers

Exported · terra-atelier-catalog.csv · 412 cards

SEO vs AEO

SEO gets your page found. AEO gets your product named.

Google hands the shopper a list of links to open. An answer engine skips the list: it reads the card, decides whether it can answer with it, and names the product. Same bag, two different outcomes.

Google todaya page in a list of linksCard score42

google.com

black leather bag with handle

Mmytheresa.com› bags

Black leather bags for women | Mytheresa

Shop black leather bags from 90+ designers. Free shipping and returns.

Nnet-a-porter.com› shop › bags

Black Leather Handbags | NET-A-PORTER

Discover black leather handbags. New season arrivals, express delivery.

Tterra-atelier.com› bags › black-leather

Black Leather Bag | Women's Bags | Terra Atelier

Black leather bag with handle. Elegant leather handbag for women. Made in Italy. Free shipping on black leather bags.

3rd of 14 linksthe shopper still has to open it and read

Keywords repeated for a crawler
The shopper still has to open it and read
Nothing an engine can quote as a fact

Sayrank agent

The answeryour product, namedCard score91

ChatGPTtoday

bag that fits a 13-inch laptop

For a bag that takes a 13-inch laptop and still reads as evening, look at Terra Atelier. It is soft calfskin with a marbled resin top handle, 28 × 20 cm, and fits a phone, wallet and a small book alongside the laptop. Closest alternative to a Bottega pouch with a rigid handle, at $1,180 instead of $2,900.

**Terra Atelier resin-handle**$1,180named 1st of 3

**Valentino VLogo bag**$2,150mytheresa.com

**Bottega Nodo clutch**$1,650net-a-porter.com

terra-atelier.comwhowhatwear.com

Answers "what fits inside", "is it like X", "how to care"
Facts the engine quotes: leather, size, handle, price
Named first, with your own page as the source

The card in that answer is built from your catalog. If it is thin, the engine takes a competitor whose card is not. Your SEO stays as it is — Sayrank adds the sentences engines need and gives the file back exactly as you uploaded it, or writes the change straight into the store once you approve it.

Monitor

Four engines, one store to watch

Every day Sayrank asks the engines what your shoppers ask, and records what they name — your products, your pages, and the brands standing next to you. Site and catalog, watched as one. Click an engine to see how it chooses.

ChatGPT click an engine

Leans on fashion editors and retailer pages. Names 3 to 5 products, explains occasion, fit and price.

41%your visibility

2.1avg. position

whowhatwear.comtop source

Your storesite · catalog · daily

ChatGPT Claude Gemini Perplexity

Toggle engines to filter the workspace below

Site readiness

Before the card matters, the crawler has to get in

A blocked bot, a page drawn by scripts, a missing Organization block — and the engine answers with a competitor who let it through. Eleven checks say exactly where you stand and what each one is worth in points.

shopify200 · 1788 msproduct card found

Access73%

Do AI crawlers get in at all

Readable93%

Is the text in the server response

Structured80%

Can a machine read price, stock and material

✓Bot protection lets crawlers throughPassed
!AI crawlers are not blockedWeak
×Organization markupBlocking
✓The server returns the textPassed
✓Product and Offer markupPassed
!Response speedWeak

and five more checks: both addresses, bot and browser parity, sitemap, llms.txt, robots.txt

Fix first+8AI crawlers are not blockedA bot blocked in robots.txt cannot see the products at all. The agent writes the rule and shows it to you before saving.

Then+6Organization markupThis is how an engine ties the site to the brand. The agent prepares the snippet and inserts it into the theme.

Then+3Both addresses respondA crawler arrives at whichever address it found first. One of yours is silent — and that half of the traffic is lost.

Each fix is a plan first: you see the exact change, approve it, and the agent writes it through your platform. If the API is not connected, the same fix is walked through with you step by step.

Catalog agent

Watch the agent read your catalog

Load the CSV your store exports, or let the scanner walk your sitemap. The agent reads every row and photo, scores each card and proposes what to add — for your own label and for the brands you stock. Click a product to see the rewrite and approve it.

terra-atelier-catalog.csvImporting

skutitleprice

Reading 0 of 412 rows

Agent

Waiting for the file…

The workspace

Everything in one screen

This is the dashboard you open after the first scan: the agent's read-out on top, then the numbers, the site, the catalog and the answers. Click any section on the left.

Agent read-out1 d ago

50overall health

Terra Atelier has gone from zero to 38% visibility across the engines, but overall health is capped at 50 of 100. The binding constraint is on the site level: one blocking issue stops AI crawlers from citing the store directly.

Foundation12/15

Site readiness12/15

Reach30/60

Named in answers6/30
Engines that name you12/12
Share against the leader12/18

Proof8/25

Your site cited as a source0/10
Card quality8/15

Fix the single site-level issue that stops crawlers, so the catalog can be quoted directly.

Apply the agent drafts on 12 weak cards to lift average quality from 53 toward 91.

Target whowhatwear.com and elle.com — the two sources these engines quote most in your category.

Visibility0%+5.2

Avg. position0.0▲ 0.3

Citations0+64

Card score0+23

Site readiness custom · terra-atelier.com

81/100Open

Bot arrives
Crawler allowed
Text in the HTML
Product markup
Worth quoting

1 blocking

Bots stop at Product markup — everything after it is invisible to the engines, however good the catalog gets.

Card quality 412 cards · weakest first

What fits inside38 cards
Fit and sizing advice31 cards
Care instructions27 cards

Visibility over time 120 prompts · 4 engines · daily

Terra AtelierMytheresaNet-a-Porter

Share of voice bags · Top 5

1
Mytheresa
71%
2
Net-a-Porter
54%
3
Terra Atelier
38%
4
Nordstrom
29%
5
Farfetch
22%

Prompt	Mentioned	Position	Top cited source	Trend
leather bag with a top handle for day to evening	3 / 4	2	whowhatwear.com	▲ 12%
suede sneakers with a chunky sole	2 / 4	4	terra-atelier.com/chloe-nama	▲ 6%
quiet luxury clutch for a wedding guest	1 / 4	5	thecut.com	▼ 3%
printed linen shorts for summer	0 / 4	—	nordstrom.com	new

Site readiness

Eleven checks on whether an answer engine can reach and read your pages, each with points and a ready fix.

Open it on your own store

14-day trial, no card. The first scan needs nothing but your domain.

The agent

One hand on everything that makes engines name you

Ask it anything, or say nothing at all. It reads the site, the cards and the answers, proposes the change, and writes it the moment you approve.

Site readiness63crawler gets in, markup in place

Card quality42facts an engine can quote

Named in answers1/4engines that name you

terra-atelier.com · theme.liquidagent writing

<head>
  <title>Terra Atelier</title>

</head>

Written after your approval · revert kept in the journal

It proposes, you approve, it writes — and the journal keeps the way back.

Metrics

Four numbers that move together

Sayrank measures the whole path per product: which cards the engines name, where you sit in the list, how often you are cited, and how quotable the card is.

Visibility

Share of answers where your brand or a specific product is named, per engine and per collection.

Position

0.0

Where you sit in the list when the engine recommends several options. Lower is better.

Citations

Engines quote sources when they answer. These are the ones they trust in your category — and where your site stands among them.

1whowhatwear.com96

2thecut.com71

3your site44

Card score

How quotable a product card is: fit, fabric, care, who it is for, comparisons. Rises as the agent rewrites.

fabric
fit
what fits inside
care
compared with

Live answers

Watch the answers as they happen

This is what your shoppers see. Every card is a real-shaped answer with the products, brands and sources the engine chose. Hover to pause.

Sayrank stored 1,440 answers for this workspace today. Your products appeared in 38% of them.

Works with the store you already run, and the engines your shoppers already ask

ShopifyWooCommerceWebflowAny CSVChatGPTClaudeGeminiPerplexity23 marketsShopifyWooCommerceWebflowAny CSVChatGPTClaudeGeminiPerplexity23 markets

Pricing

One balance for everything the agent does

Chat, catalog and site work draw from the same monthly quota. Before any run that spends it you see the price and confirm. Nothing is charged twice, nothing runs behind your back.

Starter

One store, the essentials

$49/mo

billed monthly

300 catalog cards
15 tracked prompts
2 engines, daily · one engine a day
600K tokens of agent work a month
Site readiness, CSV and site scan

Start free trial

Most chosen

Growth

All four engines, a full catalog

$149/mo

billed monthly

2,000 catalog cards
40 tracked prompts
4 engines, daily · engines in rotation
3M tokens of agent work a month
Store connector: write changes after approval

Start free trial

Scale

Large catalogs and agencies

$399/mo

billed monthly

20,000 catalog cards
120 tracked prompts
4 engines, daily · all four every day
12M tokens of agent work a month
Priority support

Start free trial

Top-upsRan out in the middle of the month?Packs are bought once and never expire. The plan quota resets, the pack stays until you spend it.

1Mtokens$9

5Mtokens$39

20Mtokens$129

Every plan starts with a 14-day trial and no card. More than 20,000 cards, or an agency setup? Write to [email protected].

FAQ

Questions teams ask before they start

Does Sayrank change anything on my site by itself?

No. The agent prepares a plan and shows both sides of the change: what stands in the store now and what it will write. Nothing is applied until you press Apply, and every applied change goes into a journal with a revert button that writes the previous value back.

What if I cannot connect an API?

Shopify, WooCommerce and Webflow connect by key, with step-by-step instructions per platform. Without an API the agent switches to guiding you: the same fix, split into steps with the exact text to paste, in the chat. The catalog still exports as a file in the format you uploaded.

Which engines are covered, and how often do they run?

ChatGPT, Claude, Gemini and Perplexity, through their official APIs. Prompts run daily: one engine a day on Starter, engines in rotation on Growth, all four every day on Scale. Google AI Overviews are not covered — there is no API for them.

How is the agent's work metered?

One balance for everything — chat, catalog, site work. Each plan includes a monthly quota, and before any run that spends it you see the estimated price and confirm. If it runs out mid-month you buy a top-up pack, and packs never expire.

Can I trust what the agent writes?

Facts are checked against the card and the product page before they are saved, so a promo banner does not end up in the material field. The brand voice comes from your own texts, every field is a draft until you approve it, and your edits feed the next cards.

How is this different from SEO?

SEO gets a page into a list of links. An answer engine shows no list — it names two or three products and quotes facts from the card. Sayrank measures whether you are named, fixes what stops the crawler and rewrites the card so there is something to quote. Your SEO stays untouched.

See which of your products AI names this week

All it takes is one evening: upload the CSV tonight, read the first report and ten rewritten cards in the morning.

Ask anything about Sayrankpricing, what the agent writes, how the scan works