This is almost certainly an auto-generated email. The system is probably set up to automatically delete inactive submissions after some time. I'm sure that no-one is actually checking or interacting with this system.
Journals (or at least the people who designed their back end systems) don't want empty or partial submissions to fill up. It sounds crazy but over years and years, that's a lot of wasted storage and database bloat. It just isn't good practice to leave those sorts of things lying around from a systems perspective - it's messy. Whether it's old inactive accounts, or partially finished submissions, they need to be cleaned up.
There is also the customer service aspect, you want people to remember to actually submit. I'm sure there are plenty of people who think they finished a submission but didn't. Or start a submission and forget about it.
Then there is the financial aspect. They make money off of publishing - whether through APCs or through subscriptions. For most journals (especially open access), high volumes of submissions means more papers published which means more money. This is even true for journals with low acceptance rates and presumably high standards. They they want those manuscripts before their competitors. That's how you nab the "high impact" stuff.
It isn't any different than an online store emailing you to "check out" when you left an item in your cart. Or any other online thing reminding you to finished whatever it is you started or come back to an unused social media account. I'm sure there is market data that shows some % of those annoying emails do convert people to active users/customers/whatever.
As for why you're getting these emails after only a couple of weeks, well someone decided that that was the timeline the journal wanted. They probably paid a consulting company for some market data that showed x emails over y weeks results in z% of dead submissions being converted.