Has AI fixed customer service?

“Representative!”

“Representative!”​

When I think about my customer service experiences, I think about that word. Along with that word comes stress and frustration.​

I’ve yelled it into the phone too many times to count. I’ve lost patience listening to menu items, which seem to have always changed (so listen carefully!). I’ve rolled the dice and pressed “0” right away, hoping that it would get me to a human. I’ve waited on hold, been connected to a human, explained my issue and been hung up on, left to start from scratch.

I have pretty low expectations for customer service. I also have some insight into why they tend to be bad.​

Early in my career, I somehow became responsible for setting up customer service for an online product. I was hardly qualified for the job, but I figured it out as best I could. I learned that there was a specific cost associated with every type of customer service interaction. The most expensive was speaking with a human. The cheapest was a self service option contained in the website FAQ’s. In between were things like email support, social media support and early versions of online chat bots. The goal was to solve as many customer issues as quickly as we could at the lowest cost possible. Avoiding the high cost interactions and pushing people to the low cost ones even had a name. It was called “deflection”.​

We regularly looked at reports to see what new levers we could pull to “deflect” issues and keep costs down. Occasionally we looked at data related to customer satisfaction, but it rarely factored into any of our decisions. It was all about the money.​

Eventually I moved out of that role and returned to my role as a frustrated customer. Having been on the other side of the curtain, I felt empathy for the people managing customer support experiences. I pictured someone in an office, staring at a spreadsheet, trying to figure out how to make me stop yelling “representative!”.​

Since then, Artificial Intelligence has grown and become more powerful. That growth gave me hope that soon I’d be able to get answers quickly. No more phone menu options or FAQs. Just a friendly, knowledgeable chat bot ready to help me out any time of day or night.​

But that didn’t happen, at least it hasn’t yet. Aside from the most basic issues, I couldn’t get answers, many times being served up links to FAQ’s. I found myself typing “Representative!” instead of yelling it.​

One day, I tried a different approach for a problem I was having.​

I use email marketing software to manage this newsletter. I use a separate platform to manage my website. A newsletter sign up on my website was stored in a list managed by my website platform. Adding that person to my newsletter platform had to be done manually. Eventually the manual effort was enough for me to seek out a way to connect the two. I started by pouring through the FAQ’s and help sections of both platforms. I couldn’t find what I needed. There was no phone number to call and email just felt like a black hole. I was spending lots of time on my problem and getting nowhere.​

At that time I was using AI to help me with all kinds of tasks. I thought “Why am I not using it for this???”

I typed in my situation and desired outcome. It returned the exact steps I needed. I followed them and minutes later my problem was solved. Since then, AI has become my customer support tool for just about everything. I only go to a company’s customer service for issues AI wouldn’t be able to handle, like checking on my specific account.​

Why can’t companies make it this easy?

I think part of it is that customer service is generally seen as a cost center, not a revenue driver. My guess is that customer service chat bots aren’t a big part of a company’s AI investments. (If you have examples where CS is seen as a revenue driver, please share! I’m curious.)​

Another factor is complexity and the number of potential issues. A proprietary chat bot focuses on issues mainly within a company’s product. Many issues involve combinations of multiple products from multiple companies, like my newsletter sign up issue. The information is out there, just not in one place. Some of it may be in blogs or review sites. One company might have the info and another might not. Organizing all of that is a lot for a cost center and probably not high on the priority list. For AI, it’s easy. It can piece all of that together into a solution.​

I think this is great. I just need my problems solved. How that gets done doesn’t really matter to me.

Thanks to AI, my customer service experience is so much better now. I’m hopeful again, but it didn’t play out in the way I expected.​

What about you? Have you had good or bad customer service experiences? Are you hopeful?