As someone who's worked in customer service: you would not believe the number of people who contact customer service with "I've lost my password", even though there is a "lost password" link right on the login page. These people are helped by the service agent following a script and clicking the link on their behalf. This is what the AI will replace.
For many businesses it's not profitable to serve customers whose issues make them fall outside the golden path of standard procedures: an average call center employee can cost $30 for a complex customer question, if you add up all cost associated from the employees involved, obliterating any margin from that customer. I doubt that AI will change this much. So if you have a question that someone with HN-level IQ cannot resolve themselves then customer service cannot profitably help you. The best you can hope for is that someone will create a ticket that someone else will look at at some point if a lot of similar tickets accumulate.
>an average call center employee can cost $30 for a complex customer question
How? I've worked in a call center where big companies outsource their customer support in Eastern Europe, and we were paid 3-4 euro an hour. Even if a "complex customer question" takes an hour (which doesn't happen often at all), what sort of overhead can there be?
> So if you have a question that someone with HN-level IQ cannot resolve themselves then customer service cannot profitably help you.
How is it a question of intellect? Customer service might have access to more information not in the public, might have access to other instances of the same problem where they had to find out how to interpret the rules and disambiguate, and most importantly, they usually have the power to take action (untick a box somewhere, offer a refund, whatever).
And how would that be profitable long term? Just by virtue of the fact that in a market with actual competition, leaving your customer dissatisfied is not profitable.
> outsource their customer support in Eastern Europe
I worked in Europe where it was not acceptable to address the customer in broken English. Easily $15 an hour. But what you are paid is not what it costs the business, (desk, training, IT support, part of a manager, etc), that can easily be double what they are paid. For anything outside the script someone with supervisor access and the authority to make high level decisions needs to get involved, those people cost even more. Some customers call several times for a complex question. $30 was just the average for complex queries.
> Customer service might have access to more information not in the public, might have access to other instances of the same problem where they had to find out how to interpret the rules and disambiguate, and most importantly, they usually have the power to take action
Most of these instances are much cheaper to solve by fixing the UI to allow the customer to do it themselves, if it isn't then it's usually cheaper to lose the customer. And you do not want to make something like "refunds" too easy.
> leaving your customer dissatisfied is not profitable
I'm sorry to say it, but dissatisfying customers that are expensive to service is often exactly what you need to do to maximise profits, as long as you manage to avoid damage to your brand. It's even better to encourage those customers to go to your competitor. What may "feel right" to you, is not what "feels right" to an MBA. Engineers tend to underestimate that.
I've also worked in a highly technical customer service role and you are absolutely right. 95% of our time was spent on 5% of our customers.
Even though they spend millions of dollars a year on support, they easily cost us more than that to help them. Simple issues that we had documented solutions for (and explained to them dozens of times preciously) were just as critical as outages to them.
We had to throttle the support we gave them in order to provide a better experience for everyone else and try to train them to be better customers.
IMHO, this is the main reason L1 support "sucks" so much. Having been a support guy in a previous life, nothing feels worse than skipping the annoying trivial question (e.g did you install the software after diwnloading it?) only to find out after 25 minutes of debugging that they, in fact, did not actually install the software. That particular call started with "all I see are codes" as the opening line. Had I asked if they'd installed the web application on their server instead of extracting random .js and .aspx files and double-clicking on them in Explorer, I could have saved everyone a lot of time.
For many businesses it's not profitable to serve customers whose issues make them fall outside the golden path of standard procedures: an average call center employee can cost $30 for a complex customer question, if you add up all cost associated from the employees involved, obliterating any margin from that customer. I doubt that AI will change this much. So if you have a question that someone with HN-level IQ cannot resolve themselves then customer service cannot profitably help you. The best you can hope for is that someone will create a ticket that someone else will look at at some point if a lot of similar tickets accumulate.