As @callebtc teaches, every business should issue ecash for its services, and people would buy their services by atomically swapping LN or sat-denominated ecash with service-denominated ecash. This isn't entirely realistic though. First, the government will kill such a business on sight. Second, the services are not necessarily fungible: for Uber rides it won't work.

This is a very good question. Westerners have strong cultural bias against haggling. However, haggling isn't inherently harmful, it's just very inefficient as practiced in an Eastern bazaar. The "dynamic prices" are basically haggling, except the customer is pretty helpless.

Here's what haggling is for. You come to the bazaar and see a thingumbob that you want to buy as long as the price is lower than 2 Msat, but you shouldn't say this out loud. The vendor wants to sell it as long as the price is higher than 1 Msat, but won't say this out loud. If you have a good haggling strategy, you might pay only 1.01 Msat, but if the seller has a good strategy, you might pay 1.99 Msat instead. Ideally a haggling protocol would converge to a point with a predefined split of extracted value, for example, if we target equal split, than an ideal haggling protocol would end up with the price of 1.5 Msat. We don't have that though.

Generally if the valuations are known, then the winner would be the party that is able to bind itself the strongest, thus disempowering itself the most. So your best strategy would be something like "I swear in the name of Satoshi that I will not pay more than 1.01 Msat for that thing!". If it's credible, then you win. Proverbial bridge burning refers exactly to the same dynamic: to win you must cut your own options as much as possible, which historically sometimes led to actually burning your only path to safety.

A couple of easy puzzles to round this up: