As Lewisham puts it, structure to maximize enjoyment of their intrinsic games is the best. Ideally, find out directly from the employees (or by being them for a while) what intrinsic games they play for their tedious tasks, and then try to build structure around making it more fun -- there's essentially 4 parts to it - holistic goals, performance stats, realtime feedback (these 3 are mentioned by example by Lewisham), and the hardest-to-implement, fun interactions.
The last is complicated, but for example, it's the reason why a clickety mechanical keyboard may be more fun to type on than one of those polymer rollable keyboards, if you have an employee whose job is to type all day [you and me probably]. Sometimes workers find ways to turn their interactions with their work fun on their own (like those in [1]), but for the less intrinsically motivated, employers facilitating it can have a huge effect.
> if we assume that people will go looking for intrinsic motivation in even the most tedious jobs
How safely can we assume that? Is there any research on it?