Skip to content
Runner AI
Automate, test, and create
English
Esc
navigateopen⌘Jpreview
On this page

Simulations

Choose a simulation mode, review its inputs, and follow generated shopper or campaign predictions.

Simulation lets you send generated shopper personas through a live store journey or ask them to evaluate one public page.

Entry path

Open a store project → MeasureSimulation.

Shopper and page simulations require publication before a run. This setup screen does not contain completed journeys or results.

Shopper and page simulations require publication before a run. This setup screen does not contain completed journeys or results.

Before you start

  • For Shopper behavior and Page stimulus audit, publish the store and confirm its public pages load.
  • Campaign prediction, when available, uses campaign inputs and does not require a published storefront.
  • Review the current plan or usage notice before launching paid work.
  • Choose a real storefront question and one target page.
  • Treat generated shoppers as directional testing, not real customer demand.

Modes

Mode When to use it What Runner does after launch
Shopper behavior You want to observe navigation, decisions, and journey friction. Starts multiple shopper journeys on the selected live page and prepares funnel, journey, outcome, and recommendation views.
Page stimulus audit You want focused reactions to one page. Collects generated shopper evaluations for purchase intent, segments, and friction on that page.

Controls and click results

Control What happens after you select it
Mode selector Chooses Shopper behavior, Page stimulus audit, or Campaign prediction when that mode is enabled.
Persona guidance Adds optional audience direction, such as price sensitivity or shipping concerns.
Target page Opens discovered public storefront pages such as Home, Collection, Product, Cart, Checkout, or a custom page.
Device viewport Chooses mobile only, desktop only, or a weighted mix for shopper behavior.
Panel size Changes how many generated shopper journeys the run requests within the allowed range.
Launch simulation Starts the configured run and opens its run page. It is disabled until a target page is selected.
Recent simulation card Opens an earlier run and its current status or results.
View all Expands the recent-run list when more runs exist.
Retry Reloads target pages or runs after a visible failure.

Launch a focused simulation

  1. Open Simulation.
  2. Select Shopper behavior or Page stimulus audit.
  3. Add concise persona guidance only when a specific audience matters.
  4. Select Target page and choose a discovered live page.
  5. For shopper behavior, choose the device mix and panel size.
  6. Review the current usage notice.
  7. Select Launch simulation once.
  8. Follow the run page until it completes or fails.

Troubleshooting

State Meaning and next step
Publish wall No successful live deployment is available. Publish and verify the store first.
No target pages Runner could not discover public pages. Open the live store and retry page loading.
No simulations yet Configure the panel and target to create the first run.
Queued or pending The run was accepted but has not started all work. Wait rather than launching a duplicate.
Running Shopper work is still in progress; results can remain incomplete.
Failed Open the error shown on the run and start a new run only after correcting the target or temporary failure.

Campaign prediction, when available

This mode asks generated audience members to evaluate a campaign rather than navigate a published store.

  1. Open MeasureSimulation and choose Campaign prediction if shown.
  2. Review the campaign inputs requested by the form.
  3. Add the campaign you want evaluated and read the current usage notice.
  4. Start the prediction only when the inputs and requested spend are correct.
  5. Open the resulting workspace and inspect the available Society, Chat, and Report views.

Checkpoint: A completed report is a generated prediction for the supplied campaign. It does not launch an advertisement or measure real campaign performance. If the mode is absent, use an available mode or check Feature availability.

Last updated on September 5, 2026

Was this page helpful?