AI should remove friction, not agents, from contact centers
No, AI will not fully replace contact center agents. Human, voice-based customer conversations that are transcribed are becoming a data gold mine for organizations.
In 2025, roughly 55,000 job cuts were linked, at least in part, to the proliferation of artificial intelligence.
Companies including Meta, Amazon and IBM have made staff reductions citing AI-driven organizational efficiency. More broadly, Google, Microsoft and Salesforce have grown increasingly explicit that AI can enable them to operate with slimmer teams.
In some instances, companies have gone further, restructuring entire job functions around AI and eliminating roles, which raises bigger questions: Could AI replace contact center agents? And, if so, is that necessarily a bad thing?
Replacing human agents is 'a mistake'
Firing employees to replace them with AI -- especially in the contact center space -- is "a mistake," said Jeff Pulver, who leads The vCon Foundation, a virtualized conversations group. Pulver is a VoIP pioneer whose advocacy and efforts on the regulatory front paved the way for modern internet-based communications.
"The mistake some companies will make is thinking the goal of AI is to remove humans from the experience. The real opportunity is to remove friction from the experience," Pulver said.
It is a widely accepted notion that AI is supposed to cut customer service costs. Yet Gartner forecasters note otherwise. They say that by 2030, costs per resolution for generative AI will exceed $3, which could exceed the cost of some outsourced agents.
New human role in the contact center
Humans are expected to step into more of an advisory role, according to contact center experts at The Office Gurus, a provider of call center services. Human agents would focus on empathy and managerial roles, while AI performs daily, repetitive tasks like ticketing and notetaking. Jaimie Bell, vice president of client solutions at The Office Gurus, described the transition as AI takes a more prominent role in the customer experience.
"It's a managed transition, and that's by design," she said. "AI handles the tactical aspects: the knowledge, the process, the procedure. But humans focus on the actual conversation. AI strengthens the human element by removing distractions, not replacing the people who deliver empathy."
Brian Peterson, co-founder and CTO of Dialpad, described a broader shift in responsibilities, from human agents doing the work to orchestrating it. Instead of handling every interaction themselves, agents increasingly guide AI systems using their experience and judgment, turning frontline knowledge into structured workflows.
According to Bell, The Office Gurus saw measurable gains in average handle time, which dropped by 45 seconds, allowing agents to manage 27% more volume without adding head count. Customer satisfaction scores also rose by as much as 25%, nixing the need to hire additional human agents.
The moment a customer needs reassurance, judgment or accountability, the human voice becomes the brand.
Jeff PulverChief evangelist officer, The vCon Foundation
Contact centers have long been under pressure to do more with fewer resources. Paired with high turnover rates, rising customer expectations and the stress of handling frustrated callers, these environments are often complex to manage and even more taxing to staff.
"Customers rarely contact a support center because things are going well. They call when something is broken, confusing or urgent," Pulver said. "Automation can resolve many routine requests quickly. But the moment a customer needs reassurance, judgment or accountability, the human voice becomes the brand."
AI reality and points of tension
Businesses care about efficiency and cost, and AI is already changing how contact centers operate. This notion aligns with what Peterson sees in the market. Most C-suite leaders, he said, are not trying to cut head count, but rather "stop the bleeding" from constant Tier 1 agent churn. By offloading repetitive work to AI, companies aim to retain experienced agents and reduce burnout.
For many leaders, the focus has shifted from whether AI will replace humans and toward where it makes sense to deploy it. Gartner found that more than 80% of organizations expect some reduction in agent head count over the next 18 months. Most reductions are gradual and not solely driven by AI, Gartner noted.
Often, AI does not free up time for workers, according to a Harvard Business Review article. The promise is that AI will lighten workloads, but it has not consistently delivered time savings. This effect compounds when teams have to navigate fragmented systems of record like a CRM system.
Some companies face integration challenges, while others see real gains in productivity and agent retention when AI is used as a tool to assist rather than replace agents.
Vendors are eager to show AI as transformational. But, for buyers, what matters most is whether these tools solve day-to-day problems inside the contact center. Here is where Pulver said humans have a leg up if they use voice communication.
"Voice remains the most natural interface humans have. When a situation becomes emotional, complicated or urgent, people reach for the phone. That will not change. What is changing is what happens to the conversation," he said.
The evolution of the contact center
For decades, communications networks carried conversations but had no memory of them. A call happened, people spoke and when the call ended, the insight disappeared. Businesses have millions of customer conversations every day but learn little from them. "That era is ending," Pulver said.
When conversations can be transcribed, structured and analyzed, the interaction is no longer a fleeting event and starts becoming a signal the organization can respond to.
"Communications systems are finally gaining memory," Pulver added. "At that point, the contact center stops looking like a cost center and starts looking like a listening system for the entire company. Every conversation becomes a source of operational insight about what customers want, where processes break and how products perform in the real world."
The AI vision in the contact center
A spike in billing complaints can now be flagged in real time. With the help of AI, this speed gives product teams immediate feedback instead of waiting weeks for reports. AI can also handle much of that initial interaction, or information gathering, assisting agents when they are requested or required to resolve issues.
Some companies see productivity gains and agent retention when AI is used to assist rather than replace agents.
"When companies start learning from their conversations, the organization becomes more adaptive. Executives who understand this shift begin to see the contact center differently," Pulver said. "Instead of measuring it only by cost per call or average handle time, they start asking a different question: What are our customers actually telling us?"
The next step is conversations themselves start to take on structure, he added.
Instead of disappearing once they end, conversations are a persistent record that systems can reference, analyze and learn from over time. At that point, Pulver noted, the economics of communication change. The call is no longer just a moment between two people.
"It becomes part of the memory of the organization," he said. "The companies that learn from their conversations will improve faster than the ones that let those conversations disappear. We're entering a period where conversations become part of the intelligence of the network. The network used to carry conversations. Now it will help organizations learn from them."
As companies adapt to a world where AI is everywhere, the challenge is learning how humans and machines can work together. Only time will tell if total human replacement will become a reality as AI advancements persist.
Moshe Beauford is a writer with nearly a decade of experience covering enterprise technology, including AI, unified communications and customer experience.
