Honestly the only thing I see current level AI reliably doing without being used as an assistive tool is grunt info work. For example a lot of T1 customer service positions can moreorless be replaced out with the current level of LLM’s that we have. Many T1 support roles consist almost entirely of searching the current customers issue, copy/pasting a boilerplate solution list of what may fix it, asking “did that work?” and if not escalating to the next tier. Hallucinations at this level won’t have a very big impact outside of annoying the customer and the t2 when it gets escalated because it failed to fix the issue. Said system shouldn’t have control over anything, it should strictly be information based. Anything management wise or financial wise or general output of merchandise should not be using these technologies standalone, at most it should be an assistive tool to a human in that position.
Idk if I am a special case or part of the majority, but when I resort to calling a support line, it’s because I have exhausted the options available to me via technology, and I need a human to take responsibility. So there has literally never been a time when a purely informational AI support bot has been useful to me.
I’m the same way here, being said, if that’s the case a T1 wouldn’t be useful to you anyway, as they just copy/paste the simple solutions you have likely already tried. So really no harm is done in this circumstance.
Being an assistant can make it a replacement. If 1 AI assisted worker can do the job of 1.25 not AI assisted workers you can drop 1 in every 5 workers and still complete the same amount of work (numbers made up).
for rewording for simplicity because I read it wrong the first time. The linked article said that in their study an AI assisted developer took an average 20% longer to complete a project than the non-AI assisted dev.
This is actually quite interesting to me, granted their study pool was very small(only 16 devs), but that is an interesting data point.
Being said, this is also a different field than what I was talking about, since that moved it to development instead of T1 customer service, but the data is nice to see.
Honestly the only thing I see current level AI reliably doing without being used as an assistive tool is grunt info work. For example a lot of T1 customer service positions can moreorless be replaced out with the current level of LLM’s that we have. Many T1 support roles consist almost entirely of searching the current customers issue, copy/pasting a boilerplate solution list of what may fix it, asking “did that work?” and if not escalating to the next tier. Hallucinations at this level won’t have a very big impact outside of annoying the customer and the t2 when it gets escalated because it failed to fix the issue. Said system shouldn’t have control over anything, it should strictly be information based. Anything management wise or financial wise or general output of merchandise should not be using these technologies standalone, at most it should be an assistive tool to a human in that position.
Idk if I am a special case or part of the majority, but when I resort to calling a support line, it’s because I have exhausted the options available to me via technology, and I need a human to take responsibility. So there has literally never been a time when a purely informational AI support bot has been useful to me.
I’m the same way here, being said, if that’s the case a T1 wouldn’t be useful to you anyway, as they just copy/paste the simple solutions you have likely already tried. So really no harm is done in this circumstance.
Most T1 customer service I deal with is already useless AI garbage.
Being an assistant can make it a replacement. If 1 AI assisted worker can do the job of 1.25 not AI assisted workers you can drop 1 in every 5 workers and still complete the same amount of work (numbers made up).
Currently, one AI assisted worker can do the job of ⅘ of a non-AI assisted worker (numbers not made up: https://arxiv.org/abs/2507.09089 )
for rewording for simplicity because I read it wrong the first time. The linked article said that in their study an AI assisted developer took an average 20% longer to complete a project than the non-AI assisted dev.
This is actually quite interesting to me, granted their study pool was very small(only 16 devs), but that is an interesting data point.
Being said, this is also a different field than what I was talking about, since that moved it to development instead of T1 customer service, but the data is nice to see.