Gartner has predicted that no Fortune 500 company will fully replace human customer service employees with AI by 2028.
The prediction comes as companies rush to invest heavily in AI, with the analyst firm forecasting that AI application software spending will climb to $269.7BN in 2026, after more than doubling to $172BN this year.
As AI spending accelerates, Sam Altman, CEO of OpenAI, has been among the AI executives vocal in warning that customer service jobs could be among the first to disappear as AI agents advance.
But Kathy Ross, Senior Director Analyst in the Gartner Customer Service & Support Practice, warned companies against attempting to replace human customer support entirely. She said:
AI and automation are transforming how customer service organizations serve customers, but human agents are irreplaceable when it comes to handling nuanced situations and building lasting relationships.
"We expect fewer human agents, but not completely agentless organizations."
What’s more, in another report, Gartner predicted that by 2027, 50 percent of organizations that planned to reduce their service workforce significantly by implementing AI will drop these plans.
Per the research firm, instead of aggressively pursuing staffing cuts, organizations should focus on using AI to scale support and give agents the tools they need to help them deliver more effective customer experiences.
The analyst also said that replacing humans with AI agents entirely in the future "is not only unlikely, it is also undesirable."
Start with the Basics...
Rather than rushing to pursue complete automation, Gartner advises that service and support leaders take a more measured approach that uses AI for basic, repetitive tasks. That would then free up employees to focus on higher-level tasks and human interactions that can enhance the customer service experience and, in turn, lead to revenue growth.
"Service and support leaders should leverage AI for efficiency, but not at the expense of human talent," Ross advised. "While AI excels at handling routine and well-defined problems, it often struggles with exceptions and high-risk scenarios.
