Can AI Replace People Working In Customer Service?

Krzysztof Szukieć
Stepwise
Published in
3 min readDec 6, 2019

I have talked about how AI can improve your security before and today we’ll take on the topic of customer service in the rise of new technologies. We’ve had chatbots before but they were just a means for a customer service operator to contact the customer. Now, chatbots are being driven by AI and could potentially eliminate people in customer service altogether. But can we actually replace people?

AI-based Customer Service Solutions

Today, we are not wondering if using AI in customer service is possible, we are wondering which one of the many solutions available to choose. There’s lots of AI chatbots, assistants, avatars. Even IBM has developed one called Watson Assistant. When you see companies like this not only using but also creating these solutions, you know it’s serious. On Facebook, you can also use intelligent chatbots (based on the LivePerson solution) to aid your customers with anything they need. Some solutions are even programmed to seamlessly transition the conversation with AI to a human conversation when needed. And this brings us back to the main question — can AI replace people in customer service?

The Importance Of Customer Service

Companies have learned that how they treat their customers really matters. Sometimes, customer service can really make or break a deal. Probably each one of us remembers a situation when a customer service agent was rude or not helpful. The impact this experience has on the customers is really significant. How we felt is usually what you will remember about using a product, even if the product itself was nice. If you felt bad — your memory of the company in general is not going to be pleasant.

The customer service interactions can be “measured” with urgency and emotion. Most cases can be solved via a simple chat or email, because they are low urgency and don’t carry much emotional value. But then, there’s some situations when the urgency, the emotional value or both get high. Situations like a security breach, network down or anything that’s really urgent and causes anger or other strong emotions. While AI can manage those simple cases, they don’t seem to be doing well with the urgent, emotional ones. And that’s because AI can learn and “think” quicker than humans do, but the emotional aspect of customer service is absent in AI solutions.

Human Agents Vs. AI Bots

Because of how different the customer service cases are, some companies decide to combine traditional customer service with AI-based solutions. This means that they can still satisfy a lot of people quicker and at a lower cost but also deal with the more intense situations. With that said, technologies like AI and ML are only as good as the data they learn from. And gathering quality data to train your AI takes time. That’s why, at the moment, AI bots and assistants are great with small, mundane tasks but aren’t so good with cases that require a human touch. After all, it’s that great, helpful lady at the call centre that can understand you and come up with a solution.

The Future Of Customer Service

Now, we know that AI is being used to provide customer service and has great results with certain cases. And we know that it takes time to train good AI. So it’s only logical to assume that in the future, AI will be able to take on more and more complicated and demanding cases of customer service.

But is it what we actually want? Well, according to Drift research, 62% of people don’t mind talking with a bot because it’s faster. And we expect responses to be as fast as a few seconds. That’s probably why 53% of companies that directly interact with customers plan to implement AI-based customer service solutions before 2022. With our attention span getting shorter and shorter and our expectations getting higher, AI customer service solutions are a great way for companies to cater to their customers’ needs. And, after all, as long as we get all of our questions answered — do we really care who answered them?

--

--