But, having the last resort option to speak to a human even when there is the ability to reach a human via other modalities can be useful.
There is a large "enterprise" company where the support folks always ignore everything I submit to the ticket, and then ask for all the things already provided. The only way to make progress with them is to get on a call. You are correct, it shouldn't be necessary-- but it is. And, if you do not have that last resort method, good luck. You may be stuck having to reply with the same information to a series of support folks without making progress on the case for weeks.
You seem to be implicitly assuming that the human on the phone has more power than the one behind a ticket system, but there's no reason to believe that. If the rule is that you need to provide specific information to recover your account, speaking to a human won't fix that when you don't have the information.
Clearly companies are en masse externalizing the costs of support onto users, and plainly not supporting users in fixing real problems with systems that are now required to live in society. It's at the stage where it's breaking our well functioning civil society and turning the first world into a third world kafkaesque dystopia.
I believe this is an economic race to the bottom and regulation is the answer to fixing it.
Here's the basis of an idea for how to fix the power dynamics so the market works properly
1) an ombusdman that can fine companies for user abuse of wasting user time, and redistribute the fine to the users 2) regulations of how long a company can on average keep users on hold, such as no more than 10 minutes to speak to a human, otherwise the speak to a human option is useless as the company makes it take longer than the median user can afford 3) regulations of threshold of acceptable customer complaints against a company, otherwise the company will make the service operators unable to actually fix anything as you note 4) a process with the ombusdman where when the number of complaints exceeds a threshold the company is fined and the fine distributed amongst the complaining users 5) some sanity check here to prevent user abuse, such as checking the average wait times, and confirming the customer service actually never fixes issues from a number of tests. The tests could be made in collaboration with a subset complaining customers.
That's a much, much, much bigger ask than just "speaking to a human".
It is not necessarily about power, it is about judgement. However clever or 'smart' the systems we design, there is still a vast gap between what automation can understand and what a human can understand. There are countless examples of a system only doing what it was told to do, where any human with an IQ higher than 80 would immediately recognize that the system is being incredibly dumb and wrong in this particular case.
There needs to be some last resort for the victims of those errors when there are real consequences to these mistakes.
It's hard for my not to wish you to have the kind of horrible automated ticket/service experience that makes everybody who had gone through it immediately understand the value of speaking to a human.
For instance I had an issue with my phone, I had to call support. I got redirected to 5 agents at 3 different companies before I finally started getting anywhere. Note some of those agents even gave me the wrong phone number and basically ended up taking me to some random unrelated companies. The end result was to start doing things over email (basically same as ticketing). After hours of phone calls and even longer on hold I'm finally (maybe) getting somewhere using emails. Still haven't gotten to the destination but maybe I will soon