It's very easy for someone to come up with an app to order a robotic taxi. This is probably the easiest part of the robotic taxi business.
It's so much so, that unless Uber or Lyft's service is essentially "free," a robotic taxi business will probably consider the cost of working with Uber or Lyft not really worth it and bypass them. They might also prefer the experience of users sticking with their own app in their own walled garden. The robotic taxis will probably have large stickers in them that say something like, "order your next car through us and $AVE!!!" (Just like how JetBlue encourages customers to use jetblue.com)
Or, to be quite honest: If I can actually hail a robotic Tesla taxi next year, I'd rather use Tesla's app instead of Lyft or Uber's app.
Meaningful such that it actually delivers monetarily
*edit for grammar
You'll have to expand on 'meaningful'. I can see driverless tech being used in specific use-cases between specific locations along controlled or established routes (example, moving cargo between a seaport and the train depot).
I don't see driverless being available as a general purpose consumer vehicle that you can summon to pick up your buddy and drive back to your work.
What about Waymo in the Phoenix area? Yes, that's a limited and perfect area, but still, it works very well there.
Besides patents, Uber has no way to bind customers. Yes, it's a commodity business, and I see no way around it. Users can simply use a different and cheaper car, which will inevitably exist and become popular.
And regarding patents: what will probably end up happening with self-driving patents is that there will be a bunch of different entities (everyone currently working on self-driving tech) that will cross license their IP to each other. A patent in this space is going to be good for licensing fees, but not for giving yourself a monopoly on self-driving tech.
Approval? That's another matter.
Google could build everything Uber has in a weekend once Waymo gets driverless cars working.