Customer Service

Can ChatGPT Improve Customer Service?

Customer service, or at least good customer service, has declined drastically over the past few years. The situation will likely worsen as companies lay off call-center employees and customer-care turnover remains high.

The answer for many businesses is artificial intelligence. Can ChatGPT and other forms of conversational AI solve the customer care problem? Companies in various industries are willing to give it a shot.

The Problem

In its tenth Customer Rage Survey of U.S. consumers, Virginia-based Customer Care Measurement & Consulting found in 2022 that 74% of respondents had experienced product or service problems in the past year — more than double the 32% in 1976.

According to the survey, only one in eight dissatisfied customers will recommend companies that do not resolve complaints. Among the most common complaints are long wait times on the phone.

For example, the U.S. Social Security Administration recently announced that telephone wait times to the national number might increase to 35 minutes (versus 14 minutes in 2021), and the response period for disability claims could stretch to 15 months. Years of decreased budgets and diminished staffing account for the service decline, per the SSA.

Many consumers prefer talking to a human via phone, but businesses are abandoning that method due to high costs.


For years companies have used chatbots for customer support. Voice-based chatbots are increasingly popular, aided by the widespread use of Siri, Alexa, and other virtual assistants.

AI-driven chatbots — text and voice — attempt to provide conversational, personalized responses. But the answers are often rote and do not address the questions. When asked again, the bots often repeat the initial reply. Nonetheless, a 2022 study from the technology research company Gartner projected that chatbots would be the main customer service channel for a quarter of all U.S. businesses by 2027.

When it publicly launched ChatGPT late last year, OpenAI stated the service is “designed to be able to generate human-like text and respond to questions and statements in a way that is similar to how a person would.” It quickly took hold as a panacea for seemingly every problem in every industry, with users often touting the results as better than humans.

ChatGPT’s Popularity

In February, asked 1,000 U.S. business leaders if they use or plan to use ChatGPT.

The results:

  • 49% of respondents currently use ChatGPT, and another 30% plan to.
  • 48% of respondents using ChatGPT say the technology has replaced workers.
  • 28% of respondents deploy ChatGPT for customer support.
  • 63% of respondents say ChatGPT will definitely or probably cause layoffs within five years.
  • 25% of ChatGPT users have already saved over $75,000.
  • Other popular ChatGPT uses are writing code and articles.

Goldman Sachs echoes the layoff prediction, stating that generative AI could cause up to 300 million global job losses.

What’s less known is the potential productivity gains and job creation from conversational AI. Bonnie Low-Kramen, author of the book “Staff Matters,” says AI tools lack empathy and intuition and cannot replace people for customer service.

“Humans understand the nuances of why people want things and do things,” Low-Kramen said in an interview. “Humans have deductive reasoning that ChatGPT doesn’t.” She points out that AI is not good at adapting when things go wrong; it has no Plan B. Low- Kramen sees a hybrid model of humans and AI as a better approach.

Consumers want to feel valued, and having to deal with machines for customer service is often demeaning, according to Low-Kramen. Thus turning over all customer-service tasks to ChatGPT could cause a backlash. Not long ago, executives in developed countries outsourced customer service to call centers in regions with cheap labor. But high employee turnover and poor English skills resulted in inadequate service. Many companies have brought call centers back domestically.

ChatGPT might look like a cost-saving cure-all, but it’s not likely the magic bullet that rectifies poor customer service.

Marcia Kaplan
Marcia Kaplan
Bio   •   RSS Feed