Checkout monitoring for online stores

Why checkouts break without anyone noticing

A checkout does not have to go down to stop selling. It can break at one step. A theme update moves the add to cart button. An app opens a popup over the basket. Bot protection starts refusing visitors that do not behave like a person with a mouse. A size picker stops responding to anything but a hover. The home page still loads, so every uptime check stays green.

A person who hits one of these may leave without telling you. An AI agent shopping for somebody always does: it has nobody to report the problem to. The stall library lists every stall an Ottom scan can report, and what to change for each.

What Ottom walks

Each journey takes a product down the purchase path on your live store, in this order. By default a script walks every step first, because most of what ends a purchase is mechanical. Where the script stops, an AI agent takes over that one step and tries to get past it. Before a stall is reported, an AI agent walks the store again on its own, and the further of the two walks is the result.

1. Open the product
2. Choose an option
3. Add to cart
4. Open the cart
5. Start checkout
6. Reach the payment step

What a checkout alert tells you

A report names the stage and the step where the agent stalled, shows a screenshot of the page it was looking at, and gives a fix written so you can forward it to whoever maintains the store.

One odd run does not page you. A Critical finding is announced only once two signals agree on it: two agent journeys, or one journey backed by a retrieval check. A score drop is announced only when it is more than 5 points and beyond the run to run variation. When the fix lands, a fix confirmed alert follows once the finding has been absent from two completed scans in a row. Every alert, and when it is sent.

How soon you hear

Checkout is walked by a full agent journey, every month on Coverage, every week on Monitor, every day on Defend, twice a day on Patrol. Price, stock, structured data and policy checks run every day on every paid plan. A checkout that breaks between two journeys is found by the next one, so the plan you choose sets how long that can be. Compare the plans.

On Shopify stores with the Ottom app installed, on Defend, Patrol and Enterprise, publishing a theme queues a retrieval check of the store rather than waiting for the next scheduled one, up to 1 a day. The app is optional: every other check runs without it.

What checkout monitoring with Ottom cannot see

• Payment itself. The agent never pays, so a payment provider declining cards is outside what it can test.
• Anything after the order. With no order placed, confirmation emails and fulfilment are not tested.
• Account checkouts. The agent never signs in or creates an account. A checkout that insists on one is reported as a finding rather than signed up to.