They really are. We’ve already heard more than rumblings about job automation and AI presumably hitting customer service hard in both the UK and US. According to the study that BBC’s article references, customer service reps have a 91% risk of losing their jobs before 2035.
In the past two years, we’ve seen investment starting to pick up for machine learning products in conjunction with the rise of chatbots. And we’re no longer talking seed rounds only.
We’ve also seen somewhat successful bots for very specialised use cases, like IBM’s Watson selling jackets, gifts or tires and Swelly that helps you collect opinions. Watson has also launched as a customizable customer service bot, and several other companies are promising automation in customer service through machine learning - companies like True AI and DigitalGenius. When talking to these chatbots, it’s clear that making machines understand natural language is still a challenge, but one that’s being solved rapidly.
Now it’s 2017, and with only a few years to go before finished products that seek to improve or even fully automate customer service hit the online shelf, a lot of people are (or should be) asking how realistic AI replacing customer service reps actually is. There are two parts to understanding what’s going on:
Yes. Well, probably. While still a bit premature, and therefore hard to pass judgement on, the technology is very promising. Given a long enough timeline, an AI will certainly be able to pass the Turing Test in any context and appear to be what most people would consider a human. But how about in the next 10-15 years? Probably not.
But there are other things going on here: You can deploy a bot for high volume, very specific enquires and processes that you can automate. That takes care of some of the need for humans. And you have to look at more than just AI: Automation can improve and speed up a lot of processes, which also lessens the need for humans operating business apps. It can also improve self-service, which deflects contacts and saves even more time in the contact center.
All-in-all, the tech should be able to replace most - but not all - humans that work in customer service today.
When weighing whether you want to replace humans in customer service, employers will consider savings from not having to pay salary and possibly benefits from providing 24/7 availability (machines don’t tire) and maybe more consistent and faster resolution of some types of enquires (due to automation). In the short term, meaning the next 10-15 years or so, there’ll still be plenty of interactions that a bot won’t be able to handle, and those will still have to be done by humans. Even so, it still sounds enticing to an employer: Do some things more efficiently while saving money.
So you can replace some of your workforce. But would an employer really want to, even if it was feasible? If you look at the costs, the answer must be a resounding no.
For now, only humans can build rapport with humans, and I’d say it’s likely to stay that way for those generations already born today. The uncanny valley of the mind takes care of that. Seen in that light, employers have to consider two major things as costs:
When you take into account that most buying decisions, even in B2B, are based on how the buyer feels, customer service’s function as a department that nurtures relationships with customers and clients becomes increasingly important.
Those things combined is what I believe will retain customer service reps’ jobs. But only if you take care to build lasting relationships with your customers. Fail at that, and you might see yourself replaced.