Sample audit report · SchemaCart

ExampleReal measurement, changed names. This is a real deep audit of a store that consented to it, run on 12 September 2026 under method 2026.09. The store appears here as Fernwick Maternity at example-maternity.shop, and its product names were replaced with invented ones of the same kind and length. Nothing else was altered: every score, count, timing and result below is what the audit recorded.

ai visibility report

example-maternity.shop

Deep audit · 50 pages measured · method 2026.09 · all eleven checks graded on one run

Print this page or save it as a PDF with your browser's own print command. A live report also carries a Markdown download.

Measured baseline Checked 12 September 2026

56/100

Partly readable

AI readiness39 / 75

Search basics17 / 25

4 Broken 2 Needs work 5 Good

Five checks passed, two need work and four are broken. The store lets the tracked AI crawlers in and publishes a plain shop summary, but no product page carries a complete set of product details, nothing on the 50 measured pages answers a buyer question in words, and every page whose speed could be measured was slow. This score records what the audit could observe on the day. It does not predict placement, rankings or citations.

Re-check after the fixes Dated on the day it runs

Not yet run

Re-check after the fixes: not yet run. This slot stays empty until the same eleven checks have been run again on the same store, so this page never shows a second score that nobody measured. The 56 out of 100 on the left is the only score here.

Do thisWork through the fix list below, then run the checks again. A re-check sets the two runs side by side and lists what fell as well as what rose.

What software could read, 12 September

All six tracked search and assistant crawlers reached the public store, and a plain shop summary was published, so an assistant can find out what the shop sells. Not one of the 38 measured product pages published a complete set of product details, nothing on the 50 measured pages answered a buyer question in words, and every page whose speed could be measured took far longer than the 2.5 second target. The search basics were in better shape than the AI checks: 17 of their 25 points against 39 of the AI checks' 75.

What a re-check would compare

A re-check runs the same eleven checks on the same store and sets the two runs beside each other, check by check, with the same captured inputs, the same method and the same result rule. It names what fell as well as what rose, because a theme update can undo a fix as easily as a developer can make one. Until that second run exists there is nothing to compare, and no number on this page is set against one that was never measured.

All 11 checks, worst first

Good earns full points, Needs work half, Broken none.

Product details 19 pts Broken

Not one of the 38 measured product pages published a complete set of product details. On 32 of them a description, a rating, or both were missing; on the other 6 the product code was missing as well.Do thisPublish one complete product record from the product template: name, description, image, price, currency, stock, a product code and any real reviews.

Answers on the page 9 pts Broken

Nothing answer-shaped was found on any of the 50 measured pages. No page set out a buyer question and its answer in plain words.Do thisWrite short, visible answers to the questions buyers ask, on the pages where they ask them, then publish the same questions and answers in a form software can read.

Page speed 8 pts Broken

Every page that returned a measurement took far longer than the 2.5 second target to show its main content on a phone, up to 12.7 seconds on one product page. One of the four rows could not be measured at all.Do thisAsk your developer to work through what loads before the main image and text, starting with the product page that took 12.7 seconds.

Titles and descriptions 5 pts Broken

9 of the 50 measured pages carried both a title and a description inside the guide ranges. On 37 the description was missing or empty, and on 4 more it ran long.Do thisGive every page its own title and its own description, inside the guide ranges, starting with the 37 that have none.

Headings 3 pts Needs work

12 of the 50 measured pages had exactly one main heading. The other 38 had two, so the product name competes with a second heading for the job of saying what the page is about.Do thisLeave one main heading on each product page and turn the other into a smaller one.

Image descriptions 3 pts Needs work

46 of the 50 measured pages passed the sampled image check. On the remaining 4, between a fifth and a quarter of the sampled images carried no text description at all.Do thisDescribe the images that carry no text yet on those four pages. It helps shoppers using a screen reader first, and software second.

AI assistant access 26 pts Good

All six tracked search and assistant crawlers are allowed to reach the public store, and the account, cart, checkout and order paths stayed closed, as they should be.Do thisCheck again after any change to robots.txt, your hosting rules or a security add-on. This is the check most likely to break quietly.

Shop summary 13 pts Good

The llms.txt file was there, with a heading and links, so an assistant can read what the shop sells without piecing it together from product pages.Do thisReview the file whenever the catalogue, the policies or the support pages change.

Page list (sitemap.xml) 5 pts Good

A page list was found and read at example-maternity.shop/sitemap.xml.Do thisKeep the list in step with the products you actually sell.

Secure address 5 pts Good

The insecure address redirected to the secure one on the same host in a single permanent hop, and the page it landed on loaded successfully.Do thisRe-check after any change to the domain or the security certificate.

One main web address 4 pts Good

All 50 measured pages pointed to one preferred web address on the same store, including the Chinese-language versions of the product and category pages.Do thisRe-check when new filters, sizes or tracking tags start appearing in your web addresses.

56 of 100 points earned. All eleven checks were graded on this run, so nothing was left out of the total, and the published method page holds the weights. Checked against public pages only: nothing was installed and nothing was written to the store. Each finding below carries its evidence and the technical detail for your developer.

The one row that could not be measured. The speed service timed out on one of the four pages it was asked about. A row that cannot be measured is marked inconclusive and left out of its grade rather than counted as a pass or a fail. The other three speed rows were measured, so page speed still has a grade, and no check on this run was excluded from the score.

What this report can and cannot say

Every audit sorts its evidence into the same three layers, so it is always clear which claims are measured and which are not made at all.

Layer 1

Technical facts

What the audit measured on the pages, in the product details they publish and in any crawler logs the owner supplied. These are facts about the store on the date given, and they can be measured again.

Layer 2

Observed visibility

What assistants actually said when we asked them, on the dates given. That is a dated observation with a recorded protocol, not a forecast, and it is not promised to repeat. No live sampling was run for this audit, so the answer cards below are invented examples.

Layer 3

Business outcome

Not promised. This report does not predict rankings, citations, visits or sales, and no number in it should be read as a promise of any of those.

Fictional AI answer examples

These three cards are invented, not records from an assistant. They show the shape of a recorded answer: the question, the date and whether the store was named. They are dated 4 August 2026 and are not part of the 12 September measurement. No live sampling was run for this audit.

"Maternity jeans that still fit after the baby arrives, under US$90"

Not surfaced Fictional example · Assistant A · 4 August 2026

Five shops named, none of them this store. Every shop that was named shows a price and a stock status on the product page itself.

"Does Fernwick Maternity nursing wear come in petite sizes"

Partial Fictional example · Assistant B · 4 August 2026

The store was named, but the answer said the sizing was not stated. Nothing on the measured pages answers a sizing question in words.

"Which shops sell maternity workwear for hot weather"

Named Fictional example · Assistant C · 4 August 2026

The store was named, and the answer repeated the wording of its shop summary file. It did not mention price or stock, because neither is published in a form software can read.

These are invented examples. A real audit records what each assistant said on the date shown, then asks the same questions again after the fixes. Nothing here is a ranking, and nothing predicts a future answer.

AI readiness findings

75 points available

Store-wide access and summary checks, then the product, answer and speed checks across the 50 measured pages. Open items first.

Do product pages publish machine-readable product details? 19 pts Broken

Not one of the 38 measured product pages published a complete set of product details. On 32 of them a description, a rating, or both were missing; on the other 6 the product code was missing as well. Software reading these pages has to guess which text is the product name, which number is the price and whether the item is in stock.

0 of 38 product pages complete

WHAT TO DO NEXTPublish one complete product record from the product template: name, description, image, price, currency, stock, a product code and any real reviews. Then check that the published values match what the page shows a shopper.

Product details are missing: description, rating.
https://example-maternity.shop/products/navy-dress
Product details are missing: rating.
https://example-maternity.shop/products/dress-n
Product details are missing: description, product code, rating.
https://example-maternity.shop/products/everyday-cycling-shorts

Do the pages answer common buyer questions? 9 pts Broken

Nothing answer-shaped was found on any of the 50 measured pages. No page set out a buyer question and its answer in plain words, and no page published the same questions and answers in a form software can read. For this catalogue the missing answers are the ordinary ones: sizing, delivery, returns, care and who a piece suits.

assessed: 50fail: 50

WHAT TO DO NEXTWrite short, visible answers to the questions buyers actually ask, on the product, category and help pages where they ask them. Then publish the same questions and answers in a form software can read, and check that the two say the same thing.

Do the pages load in time? 8 pts Broken

Speed was measured on the homepage and on the first three product pages the crawl reached. Every page that returned a measurement took far longer than the 2.5 second target to show its main content on a phone. An assistant that does not wait reads whatever has arrived by then.

Page	Main content shown, on a phone
Homepage /	5.0 seconds
Navy Dress /products/navy-dress	12.7 seconds
Soft Rib Flared Leg Pants Black /products/soft-rib-flared-leg-pants-black	The speed service timed out on this page, so it could not be measured.
Soft Rib Straight Leg Pants Brown /products/soft-rib-straight-leg-pants-brown	7.5 seconds

WHAT TO DO NEXTAsk your developer to work through what loads before the main image and text on these pages, starting with the product page that took 12.7 seconds. Worth 8 points.

Can tracked AI crawlers reach the public store? 26 pts Good

All six tracked search and assistant crawlers are allowed to reach the public store. The audit checked the front page and 40 private paths for every tracked crawler, and the account, cart, checkout and order paths stayed closed, as they should be.

Reported and not scored: the training crawlers GPTBot, ClaudeBot and Google-Extended are all allowed here. Those settings govern training, and for Google-Extended also grounding. They do not decide whether Google Search includes the store.

KEEPCheck again after any change to robots.txt, your hosting rules or a security add-on. This is the check most likely to break quietly.

Show the technical detail (for your developer)

robots.txt: one group naming the tracked AI crawlers
GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Claude-SearchBot,
Claude-User, PerplexityBot, Perplexity-User, Google-Extended
Allow: /  with the private paths repeated as Disallow

Is there a plain-text shop summary? 13 pts Good

The llms.txt file was there, with a heading and links, so an assistant can read what the shop sells without piecing it together from product pages.

KEEPReview the file whenever the catalogue, the policies or the support pages change.

Search basics

25 points available

The same 50 measured pages, and the same grading rule: a check takes the worst result seen on any page.

Are titles and descriptions within the guide ranges?5 ptsBroken

9 of the 50 measured pages carried both a title and a description inside the guide ranges. On 37 pages the description was missing or empty. On 4 more it ran long, one of them to 319 characters against a guide range of 50 to 160.

assessed: 50pass: 9needs work: 4fail: 37

WHAT TO DO NEXTGive every page its own title and its own description, inside the guide ranges, starting with the 37 that have none. Worth 5 points.

Does each page have exactly one main heading?3 ptsNeeds work

12 of the 50 measured pages had exactly one main heading. The other 38 had two, so the product name competes with a second heading for the job of saying what the page is about.

assessed: 50pass: 12needs work: 38

WHAT TO DO NEXTLeave one main heading on each product page and turn the other into a smaller one. A theme change, worth 1.5 points.

Do sampled images have useful text alternatives?3 ptsNeeds work

46 of the 50 measured pages passed the sampled image check. On the remaining 4, between a fifth and a quarter of the sampled images carried no text description at all. This also matters to shoppers using a screen reader.

assessed: 50pass: 46needs work: 4

WHAT TO DO NEXTDescribe the images that have no text yet on those four pages. Worth 1.5 points.

Does each page point to one preferred web address?4 ptsGood

All 50 measured pages pointed to one preferred web address on the same store, including the Chinese-language versions of the product and category pages.

Is there a parseable page list?5 ptsGood

A page list was found and read at https://example-maternity.shop/sitemap.xml.

Does HTTP redirect cleanly to HTTPS?5 ptsGood

The insecure address redirected to the secure one on the same host in a single permanent hop, and the page it landed on loaded successfully.

The fix list, in order

Six checks are carrying points this store has not earned. The order puts the points still available against the size of the job, so the fix that returns the most for the least work comes first.

How this order is set: points still at stake divided by the size of the job, counting a small job as one unit of work and a medium job as two.

Fix	Points at stake	Who does it	Effort
Publish the product name, description, image, price, currency, stock, product code, and any real reviews in a form software can read	19.0	Developer	Medium
Give every page its own title and its own description, inside the guide ranges	5.0	Owner	Small
Write clear answers to buyer questions on the right pages, then add the same questions and answers in a form software can read	9.0	Owner	Medium
Work through what loads before the main image and text on the slow pages	8.0	Developer	Medium
Leave one main heading on each product page, so the product name does not compete	1.5	Developer	Small
Describe the images that still carry no text	1.5	Theme	Small

The product details, the main heading and the image descriptions all live in the theme, so a theme update can undo them. The re-check below shows exactly that happening to image descriptions.

44 points in total, which is the gap between the 56 points earned and the 100 points assessed. Page speed is on this list because it was measured on this run, and the single row the speed service could not return was left out of the grade rather than counted either way.

One check got worse

Image descriptions

On 12 September the sampled images on 46 of the 50 measured pages carried useful text, and the check earned 1.5 of its 3 points. In this made-up re-check on 26 September a theme update replaces the product gallery and drops the short text that describes each image, so the check falls to Broken and loses all 3 points. A report that shows only improvements is not evidence.

When	Result	Points
Measured baseline, 12 September 2026	Needs work	1.5 of 3
Made-up re-check, 26 September 2026	Broken	0.0 of 3

Only the re-check row is invented. The 56/100 score, the findings and the fix list above are the 12 September measurement, and this section changes none of them.

This is the ordinary way a passing check starts failing on a store that is still being worked on, which is why every re-check lists what fell as well as what rose.

Draft artifacts

Deep audit

Review-only proposals, written from this measurement. Nothing here has been published to the store, and each file opens with the line that says so. These are excerpts; the audit hands over the whole file.

robots.txtDraft

DRAFT: review before publishing

# SchemaCart: tracked AI crawlers; keep private paths repeated in this group.
User-agent: GPTBot
User-agent: OAI-SearchBot
User-agent: ChatGPT-User
User-agent: ClaudeBot
User-agent: Claude-SearchBot
User-agent: Claude-User
User-agent: PerplexityBot
User-agent: Perplexity-User
User-agent: Google-Extended
Allow: /
Disallow: /account
Disallow: /cart/
Disallow: /checkout
Disallow: /orders

Sitemap: https://example-maternity.shop/sitemap.xml

llms.txtDraft

DRAFT: review before publishing

# Fernwick Maternity

## About

- [Store homepage](https://example-maternity.shop/)

## Products/Collections

- [Maternity Dresses](https://example-maternity.shop/collections/maternity-dresses)
- [Nursing Clothes](https://example-maternity.shop/collections/nursing-clothes)
- [Navy Dress](https://example-maternity.shop/products/navy-dress)
- [Marlow Eyelet Nursing Midi Dress](https://example-maternity.shop/products/marlow-eyelet-nursing-midi-dress)

product.jsonldDraft

Show the technical detail (for your developer)

{
  "DRAFT: review before publishing": "DRAFT: review before publishing",
  "@context": "https://schema.org",
  "@type": "Product",
  "name": "The Everyday Cotton Rib Column Dress - Short Sleeve - S",
  "description": "REVIEW: Fictional stand-in description for this sample. Replace it with the owner-reviewed store description before publishing.",
  "image": "https://example-maternity.shop/cdn/shop/files/1772193089248.jpg?v=2505740063&width=1920",
  "sku": "D-BD-NVY-143-S",
  "gtin": "D-BD-NVY-143-S",
  "offers": {
    "@type": "Offer",
    "price": "53.00",
    "priceCurrency": "SGD",
    "availability": "http://schema.org/OutOfStock"
  }
}

Per-page appendix

50 pages

Every measured page against every page-level check. Inconclusive means the check did not apply to that page, or, for the one product page the speed service timed out on, that it could not be measured.

Page	Title and description	Preferred web address	Main heading	Image text	Answers	Page speed	Product details
Homepage /	Good	Good	Good	Good	Broken	Broken	Inconclusive
Navy Dress /products/navy-dress	Broken	Good	Needs work	Good	Broken	Broken	Needs work
Soft Rib Flared Leg Pants Black /products/soft-rib-flared-leg-pants-black	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Soft Rib Straight Leg Pants Brown /products/soft-rib-straight-leg-pants-brown	Broken	Good	Needs work	Good	Broken	Broken	Needs work
Soft Rib Tank Top Oatmeal /products/soft-rib-tank-top-oatmeal	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Soft Rib Wrap Top Oatmeal /products/soft-rib-wrap-top-oatmeal	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Soft Flared Rib Dress Oatmeal /products/soft-flared-rib-dress-oatmeal	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Dress K /products/dress-k	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Dress M /products/dress-m	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Dress N /products/dress-n	Needs work	Good	Needs work	Good	Broken	Inconclusive	Needs work
Everyday Cycling Shorts /products/everyday-cycling-shorts	Broken	Good	Needs work	Good	Broken	Inconclusive	Broken
Rib Twist Tank Top Pink /products/rib-twist-tank-top-pink	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Pinspot Babydoll Top /products/pinspot-babydoll-top	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Drapefold Inspired Rib Tank Dress Indigo /products/drapefold-inspired-rib-tank-dress-indigo	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Morningside Denim Shorts /products/morningside-denim-shorts	Broken	Good	Needs work	Good	Broken	Inconclusive	Broken
Morningside Straight Leg Jeans /products/morningside-straight-leg-jeans	Broken	Good	Needs work	Needs work	Broken	Inconclusive	Broken
The Everyday Cotton Rib Column Dress Sleeveless /products/the-everyday-cotton-rib-column-dress-sleeveless	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
The Meadow Cotton Stretch Midi Dress /products/the-meadow-cotton-stretch-midi-dress	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Classic Striped Shirt Dress Indigo /products/classic-striped-shirt-dress-indigo	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Marlow Eyelet Nursing Midi Dress /products/marlow-eyelet-nursing-midi-dress	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Navy Dress /zh/products/navy-dress	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Soft Rib Flared Leg Pants Black /zh/products/soft-rib-flared-leg-pants-black	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Soft Rib Straight Leg Pants Brown /zh/products/soft-rib-straight-leg-pants-brown	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Soft Rib Tank Top Oatmeal /zh/products/soft-rib-tank-top-oatmeal	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Soft Rib Wrap Top Oatmeal /zh/products/soft-rib-wrap-top-oatmeal	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Soft Flared Rib Dress Oatmeal /zh/products/soft-flared-rib-dress-oatmeal	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Dress K /zh/products/dress-k	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Dress M /zh/products/dress-m	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Dress N /zh/products/dress-n	Needs work	Good	Needs work	Good	Broken	Inconclusive	Needs work
Everyday Cycling Shorts /zh/products/everyday-cycling-shorts	Broken	Good	Needs work	Good	Broken	Inconclusive	Broken
Rib Twist Tank Top Pink /zh/products/rib-twist-tank-top-pink	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Pinspot Babydoll Top /zh/products/pinspot-babydoll-top	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Drapefold Inspired Rib Tank Dress Indigo /zh/products/drapefold-inspired-rib-tank-dress-indigo	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Morningside Denim Shorts /zh/products/morningside-denim-shorts	Broken	Good	Needs work	Good	Broken	Inconclusive	Broken
Morningside Straight Leg Jeans /zh/products/morningside-straight-leg-jeans	Broken	Good	Needs work	Needs work	Broken	Inconclusive	Broken
The Everyday Cotton Rib Column Dress Sleeveless /zh/products/the-everyday-cotton-rib-column-dress-sleeveless	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
The Meadow Cotton Stretch Midi Dress /zh/products/the-meadow-cotton-stretch-midi-dress	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Classic Striped Shirt Dress Indigo /zh/products/classic-striped-shirt-dress-indigo	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Marlow Eyelet Nursing Midi Dress /zh/products/marlow-eyelet-nursing-midi-dress	Broken	Good	Needs work	Good	Broken	Inconclusive	Needs work
Accesories /collections/accesories	Broken	Good	Good	Needs work	Broken	Inconclusive	Inconclusive
All Products /collections/all-products	Good	Good	Good	Good	Broken	Inconclusive	Inconclusive
Best Sellers /collections/best-sellers	Needs work	Good	Good	Good	Broken	Inconclusive	Inconclusive
Bump Friendly /collections/bump-friendly	Good	Good	Good	Good	Broken	Inconclusive	Inconclusive
Everyday Staples /collections/everyday-staples	Good	Good	Good	Good	Broken	Inconclusive	Inconclusive
Maternity Bottoms /collections/maternity-bottoms	Good	Good	Good	Good	Broken	Inconclusive	Inconclusive
Maternity Dresses /collections/maternity-dresses	Good	Good	Good	Good	Broken	Inconclusive	Inconclusive
Maternity Jeans /collections/maternity-jeans	Good	Good	Good	Needs work	Broken	Inconclusive	Inconclusive
Maternity Tops /collections/maternity-tops	Good	Good	Good	Good	Broken	Inconclusive	Inconclusive
New Arrivals /collections/new-arrivals	Needs work	Good	Good	Good	Broken	Inconclusive	Inconclusive
Nursing Clothes /collections/nursing-clothes	Good	Good	Good	Good	Broken	Inconclusive	Inconclusive

Two ways forward

Watch the score while you fix it yourself, or have the fixes done with the evidence attached. The free scan is the same first step either way, and it needs no login, email or card.

Monitor your store, from $29 USD a month Book the deep audit, $490 USD

The audit measures access, content and the product details software can read. It never promises placement, rankings or citations.