Open the inventory editor
The workflow starts from a resettable inventory lab that behaves like a searchable record-management screen.
BULK RECORD UPDATES
Update Website Records from Structured Data
Apply structured price and stock changes to a browser-based inventory system while keeping every selector, field, and verification step inspectable.
WHAT THE AI WILL BUILD
A real visual workflow, not a conceptual diagram.
The AI guide supplies a precise goal, safe target page, inputs, output contract, and verification rules. Paste that prompt into Runavelo's built-in assistant, then inspect and edit every generated command before execution.
- Search for a record by SKU before editing it.
- Replace price and stock values through CDP silent input.
- Save each change and read the resulting page status.
- Separate source data from browser interaction so the input can later come from Excel, CSV, or an API.
Responsible use: adapt this pattern only for systems and data you are authorized to automate.
WORKFLOW WALKTHROUGH
How the visual workflow operates.
Each stage has a specific input, browser context, output, and verification point. Edit these commands directly when the destination system differs.
Search by stable identifier
For each input row, the SKU is placed in the search field through CDP and the matching record is selected for editing.
Replace editable values
Price and stock fields receive the values associated with the current SKU. Each field is located by a documented CSS selector.
Save the current record
A CDP silent click activates the save control without moving the physical pointer or obscuring another desktop task.
Verify before continuing
The page status is read and logged for the current SKU before the list loop advances to the next record.
EXPECTED OUTPUT
Evidence you can inspect after the run.
- Updated price and stock values for the selected demo SKUs.
- A save-status message captured after each update.
- One log entry per processed record for audit and troubleshooting.
Runavelo's runtime log connects the visible workflow line to its result. That makes missing elements, rejected updates, and unexpected page states easier to locate than a generic success message.
COMMON QUESTIONS
What to change for a real workflow.
Can this be driven by Excel rows?
Yes. The included application uses spreadsheet-like objects for a deterministic demo. Replace the list with a spreadsheet read command while preserving the same browser loop.
How should production failures be handled?
Add an explicit branch for missing records or rejected saves, log the source identifier and response, and decide whether to continue, retry, or stop based on business risk.
Copy the goal, let AI build, then inspect and run.
The prompt targets the matching test page and states the expected evidence. A reference QYA remains available for comparison with a verified implementation.