Too bad low-level support workers cannot do anything. All they will be able to do is read the AI-auto-generated text to you and answer your questions by reading from the prepared responses. You pay for having a human read to you what the computer system produced. For supporters to be cheap and exchangeable (for the business) they are bound to executing the prepared algorithm and script without option to deviate. Often they cannot even access relevant information about you, never mind updating anything in the database.

To get someone able to actually make decisions, especially when they are against a measure the system automatically put in place following its programmed or AI-.derived rules, you need to go at least two levels higher. Even "managers" often - usually? - only have digression within pre-determined possibilities and scenarios.

It's going to be like this: https://twitter.com/cnbc/status/1447916881009127430