Within high technical complexity space, they differ in significant ways with Apple or Tesla – they don't have organizational mechanisms to do deep vertically integrated problem solutioning. Even when they have academically much superior AI tech, their ability to productize those capabilities is slow and less effective due to how cross-organization collaboration works.
On the business side, they don't have a business team really. There's no MBAs scheming new pricing or bundling models nor are there people wanting to chase/beat market competitors. So, whenever they do SKU based pricing or subscription based pricing, they tend to get it less right than, say, Microsoft.
Edit: One thing to add... Google's history of 'solving highly technical problems' has really been more around scaling infrastructure, with their notable successful products being Google search, Maps, Gmail, and YouTube - with Android being an exception that rule if you'd call that a 'product'. Maps and Google assistance have some technical aspects, but nothing really out of the grasp of a modern startup. Technical ability for SaaS companies isn't really the moat it used to be.
Not any more. These days their DNA is ads.
This may be part of the problem. Google was founded by algorithm nerds, and that's now part of the DNA of the company. Apple (Tesla/Starlink/etc) was founded upon vision - what could the future look like? How do I make that happen today? - and then, somehow, to pare that back to something that's economically viable in a reasonable time frame.
That said, taking existing problems and trying to better apply algorithms to make a better product is still a thing in AI space though - and even here plenty of start ups are taking the lead.