I agree that things won't change enough in service jobs to matter by 2019. But as to this:
> The tech necessary to replace front line service workers is not there. If it was we'd all be buying one ... what happens when ... some customer needs help in unexpected ways
History does tell us that sometimes the market will throw away jobs that are becoming inconvenient to society, even if that means doing without the services that class of worker provides, or receiving worse and less convenient services.
Consider household servants.
They used to be seen as necessary for middle and upper-class households, because professional households required staggering amounts of manual labor just to keep everyone fed and clothed.
Those jobs vanished, right around the time that clothes and dish-washing machines were invented. But it's interesting to look at why.
Clothes and dish-washing machines are, frankly, not as great as servants (my feeling 'icky' about having someone else wash my clothes aside). You have to do the work yourself, and accept more cognitive load. But they win on cost -- the costs of employing a servant went up, and eventually worse service was worth it for a simpler and cheaper self-serve solution.
Could something similar happen in service work?
Well, look at the Amazon store, and compare it to a current grocery store.
Now imagine a Starbucks with touchscreen-and-voice enabled ordering, no visible employees (maybe one employee, nominally a store manager, working the fleet of coffee machines behind the screens, and doing some light sweeping in the restaurant, etc) and a communal co-working space that more or less preserves the current experience, just minus the baristas.