If you really want to go the way you've described, where you manually navigate to the page before you launch the extraction, you need to use the 'Launch New {Browser}' action anyway, but then instead of actually launching a browser, select the option to attach to an already running instance:
You can then choose the way you want to identify the tab you're attaching to:
- By title (assuming it is the same for all of your cases)
- By URL (probably would not work for your use case)
- To the foreground window (basically the active browser tab, which could work in your case)
Once done, you'll have a browser instance that the further web automation actions will be able to use.
Alternatively, I would suggest trying to automate the initial page that contains links to these incidents - open the browser, navigate to this page (assuming the main URL is static), extract the links to a list and loop through them, opening each link individually and extracting the data.
This could work, if there is an easy way for the flow to identify which incidents were already processed, if the list can contain them, too.
