Your store, through a shopper’s eyes

Find what gets in the way of buying from your store.

Build your shopping test

Meet Pulse

Mystery shop your store
with your own AI.

A first-time buyer. A delivery question. A small screen. Pick a journey and get a brief that tells your AI what to test — and what evidence to bring back.

Pulse writes the brief. It doesn’t run the audit.

You’ll need an AI assistant that can actually browse and interact with websites. Don’t have one? Use the manual checklist instead.

01 / Configure your test

Give your shopper a starting point.

No sign-up.
No automatic audit.

Store and shopping test details

Your own store, or one you’re authorised to test. Public domains only.

Be specific: include the size, colour or option you’d choose.

A town or postcode only. Don’t enter a street address or personal details.

What should your shopper focus on?

Nothing is sent or tested when you generate a brief.

Your shopping brief

Not generated

A set of instructions to review, then run in your own browser-capable AI assistant.

Enter your details, then generate a brief.

A test, not a transaction. Stop before buying, submitting checkout or creating an account. No payment details, logins, uploads, contact forms, mailing lists or CAPTCHA bypass. Add to cart only with the store owner’s authorisation. A shipping estimate may use a town or postcode only, if it doesn’t commit you to anything.

02 / Run it yourself

The brief is the beginning.

You’re in control
of the test.

  1. Review & copy

    Read the brief, check the product and confirm you’re authorised to test the store. Copy it or download it as text.

  2. Use a real browser

    Paste into your own browser-capable AI assistant. Confirm it can interact with the site. If it can’t, ask a person to follow the checklist — don’t accept an invented audit.

  3. Keep the evidence

    Ask for steps, exact URLs and screenshots if available. Separate observed problems from potential concerns and things not tested. Reproduce a finding before fixing it.

For a mobile test, the assistant must use and report an actual narrow browser viewport or mobile device. A desktop visit is not a mobile usability test.

03 / The human alternative

No browser-capable AI?
Be the shopper.

Use a fresh browser session and follow a checklist yourself. The same safety boundaries apply. Record the date, browser, viewport and exact page URLs as you go.

Capture screenshots only if you can. Keep up to five reproducible findings with a practical next action. It’s fine to find nothing, or to mark a step Not tested.

First purchase
  1. Start at the store URL. Find the exact product and variant; do not silently substitute another one.
  2. Check price, availability, product details and the information a first-time buyer needs.
  3. Find delivery costs, timing and returns information before committing. Record what remains unclear.
  4. If authorised, add the intended variant to cart. Inspect the cart and, if accessible without entering any personal information, the guest-checkout entry screen. Stop before checkout submission, login or personal-data entry.
Delivery doubts
  1. Find the exact product and variant. Locate delivery information from the product page.
  2. Check dispatch versus transit times, destination exclusions, returns and any free-shipping threshold.
  3. Use only a town or postcode in a clearly noncommitting estimator, if available. Stop if personal data or checkout submission is required.
  4. Record the known charges and what remains unknown for this location. Do not guess a total or an arrival date.
Mobile shopping
  1. Use an actual mobile browser or a browser viewport set to 390 × 844 CSS pixels. Confirm the actual viewport, browser and device/emulation context.
  2. If that capability is unavailable, mark the mobile journey Not tested; never infer mobile usability from desktop.
  3. Find the exact product and use its variant controls. Check navigation, readable details, tap targets and horizontal overflow on the narrow viewport.
  4. Find delivery and returns information. If authorised, inspect the cart on the same viewport; stop before checkout submission, login or personal-data entry.

04 / Know what to look for

Useful beats
impressive.

One concrete finding is more useful than a polished report full of guesses. You’re looking for a thing you can see, reproduce and act on — not a conversion score.

Fictional example — not a live store result. Everything in this example, including URLs, context and observations, is invented to show the format.

Fictional result / Delivery doubts

Delivery cost is unclear on the product page.

Context
Illustrative test: 14 September 2026, 10:15 AEST. Example browser, desktop viewport 1366 × 768. Linen shirt, blue / M. Fitzroy 3065.
Status
Observed — in this fictional example only.
Severity
Makes the task harder: the shopper can’t work out delivery cost before deciding whether to proceed.
Evidence
https://shop.example.com/products/linen-shirt
The imaginary page says “Shipping calculated at checkout” with no visible link to delivery rates. No screenshot is supplied.
Reproduce
Open the product URL → choose blue / M → read the delivery text beside the price → look for a link to shipping costs.
Next action
Link a clear delivery-cost guide near the price, then repeat the test to see whether the question is answered.
Not tested
Checkout, final shipping total and mobile behaviour. No order placed.

From finding to fixing

Found a sticking point?
Let’s work on it.

StewArt Media can help investigate buying-journey friction and prioritise what to test next. Bring your brief and the evidence you actually collected.

Talk to StewArt Media Learn about conversion rate optimisation