This highlights the potentially unrealistic spend on AI, where the proposal is to spend 10s of millions to save... marginal millions... that could have also been saved by a change in process with limited additional spend.

I also would assume that there are far more significant behavioral or human factors that consume the time writing those minutes, i.e. an easy spot to kill 5-10 min before opening the line for the next inbound call, but the 5-10 minute break will persist anyway.

I fully believe AI will create a lot of value and is revolutionary, especially for industries where value is hidden within data. Its the pace of value creation that stands out to me (how long til its actually useful and better and creates more value than it costs??) but the bubble factor is not ignorable on the near term.