When automated teller machines (ATMs) first showed up in the late 1970s, many predicted a steady decline in bank tellers.
I can remember working at a small community bank in 1988 and having our first ATM installed. It created a lot of discussion at my level - I was a teller at the time - about potential job replacement.
After all, ATMs could do the core transactional work like dispense cash, accept deposits, and report balances.
But that didn’t happen, for reasons I will explore in this article. Tellers are still here but in a more professional role. And coming full circle, my son just graduated college and started his first professional role as a universal banker.
...the [ATMs] took over predictable, repeatable tasks, while humans moved up the value chain.
This parable matters for call centers today as AI and chatbots will change how customer service is delivered, but they will not make human agents obsolete. Instead, they’ll shift the job toward real-time judgment, empathy, and complex decision-making where machines still fall short.
A Caution Against Linear Thinking
Economists who studied the rise of ATMs found there were two distinct outcomes.
First, automation replaced routine aspects of the teller job and reduced the number of tellers required per branch.
But second, ATMs lowered operating costs per branch and made it economical for banks to open many more branches.
The result was total teller employment did not collapse. Instead, and surprisingly, it grew for a time as the role shifted toward sales, problem-solving, and higher-touch customer service (source: Bessen, James E. “Toil and Technology”. Finance & Development, International Monetary Fund).
In other words, the machines took over predictable, repeatable tasks, while humans moved up the value chain.
Back in the day, these changes were hard to digest for the actual bank employees (me). As ATM networks expanded, less tellers per branch were required. New branches weren’t built immediately, and teller job growth plateaued for a while before it started to grow again.
Once it did start to grow, often those teller jobs morphed into personal banking - or operations roles like call centers - which took more training.
Call Center vs. Bank Teller Work
That pattern of automation substituting for repetitive tasks, while expanding or reshaping complementary job skills, is an applicable model for contact centers. It shows how organizations and markets reconfigure themselves around new capabilities.
Both bank tellers and call center agents perform a blend of routine and non-routine tasks, with automation excelling at the predictable functions such as balance inquiries, password resets, and billing questions.
In addition, both roles involve direct customer interaction, requiring employees to humanize the experience.
In my work conducting observations in contact centers, I consistently see how much on-the-spot judgment agents must use.
These employees are often making decisions with incomplete information, weighing short-term fixes against long-term customer value, and applying discretionary empathy, whether that means choosing to escalate a case or to offer a goodwill credit.
I watch agents navigate rapid task switching and highly nuanced interactions, as callers frequently present multiple issues at once or bring strong emotions into the conversation.
Ultimately, the contact center agent role is in for a makeover as AI will require the human to elevate.
These situations require agents to read tone, history, and subtle behavioral cues that technology simply can’t interpret or reason through the way a human can.
Customers choose to call precisely because they want a genuine human interaction with those who can understand, adapt, and reassure in ways automated systems are still far from replicating.
McKinsey’s recent analysis of contact centers underscores this hybrid future. Generative AI (GenAI) and automation are taking over heavy volumes of routine interactions, while organizations redesign the human role around customer advocacy.
The firm highlights case studies where human-AI combinations produce better outcomes than AI alone. And we’ve seen this in our experience.
One of our clients, a leading energy company, successfully reduced its billing call volume by around 20% and shaved up to 60 seconds off customer authentication by integrating an AI voice assistant into its back-end call workflow. The company is now planning to scale this use case across the organization.
Other customer care leaders are naturally eager to invest in this technology, too, given the promised benefits of greater efficiency and productivity in an environment that generally struggles with high agent churn and the associated costs of recruitment and training.
The upside of the technology also extends to customer and employee experience. GenAI can handle simple queries and make interactions more efficient while reducing wait times and enhancing customer satisfaction.
Agents themselves are seeing the positive effects of GenAI in their day-to-day duties, especially from reduced after-call work (ACW). AI tools can summarize issues and proposed interventions, increase agent productivity, and reduce their call times.
What This Means for Workforce Strategy
If the ATM evolution taught us anything, it’s that companies shouldn’t view AI as a simple way to cut headcount.
Instead, AI works best when it reshapes the work itself. It should automate the predictable tasks, so humans don’t get bogged down in routine work. In the contact center space, AI will augment agents by giving them real-time prompts along with instant access to the right information.
Ultimately, the contact center agent role is in for a makeover as AI will require the human to elevate. The key to their future success will be their ability to learn how to use good judgment and problem solve the most challenging scenarios.
Organizations will need to invest in training and skilling their agents to handle this higher-order work. The shift from agents performing primarily transactional tasks to acting as problem solvers who apply judgment and emotional nuance will require a far more advanced workforce strategy.
Organizations will also need to rethink the entire talent lifecycle, beginning with the screening process for this new, elevated candidate who can have higher degrees of emotional intelligence: something AI cannot do for their customers.
Training will need to move beyond transactional to emphasize advanced communication supported by continuous upskilling as tools and expectations evolve.
Supervisors will also need to be retrained to coach agents on these skills to build reasoning skills, driving customer outcomes instead of narrowly defined productivity metrics.
Together, these changes point to a broader operating model shift in which agent roles are treated as skilled knowledge work, with corresponding implications for performance management, career paths, and long-term retention.
The future of the contact center isn’t just smarter technology; it’s a more capable human workforce that knows how to interpret and action AI-generated insights.
Consider the use of AI with speech analytics in the contact center. This technology can surface the patterns, the risks, and the opportunities, but only a well-trained human can turn that information into empathy, decisions, and outcomes that build real customer value.
Closing Thought
Machines change jobs, humans change roles
As I reflect on the ATM story from my younger self, I’m reminded that it isn’t a blanket defense of every job. Automation will absolutely streamline some roles: and in certain areas it will reduce headcount.
But time and again - from bank tellers to factory teams - we’ve seen that technology tends to reshape work more than it eliminates it. I see the same future for call centers.
AI will take on repeatable, rules-based tasks, but the heart of the work will still rest with people. Only humans can rely upon good judgment to determine when to make the exception or show empathy to create better customer experiences at each point of contact.
The future of the contact center...it’s a more capable human workforce that knows how to interpret and action AI-generated insights.
Of course, transformation demands change, and most of us aren’t naturally comfortable with that. Leaders must guide teams through shifting responsibilities: and agents need to stay open to stepping into more human-centered roles.
But here’s the part I find reassuring: the tasks that matter most, like listening, understanding, decision-making, and building trust, remain deeply human. When we frame AI as a partner rather than a threat, agents can feel more secure, more valued, and genuinely excited about where their work is heading.
Case Study: JPMorgan Chase Invests in Branches and AI
Even as JPMorgan pours billions into AI infrastructure, the bank is simultaneously expanding its physical footprint. Since 2018, JPMorgan Chase has opened more than 1,000 new branches (CNBC).
Many of these branches are not “old-fashioned teller counters,” but modern “advice centers,” focused on developing relationships and creating long-term clients.
These moves underline a lesson along the same lines to the ATM story; automation rarely eliminates the need for human presence as customers may seek human interactions for confirmation, but it will reshape it.
While self-service technologies and AI can handle speed, scale, and routine execution, customers will seek human interaction for reassurance, like in moments of uncertainty where there is complexity or emotional weight. Human agents provide confirmation that the system has understood their intent.
Rather than replacing people, automation elevates the importance of human roles as validators through relationship building, which reinforces the idea that technology works best not as a substitute for human interaction, but as a complement that amplifies it.
Most tellingly, Jamie Dimon himself has publicly explained his strategy to leverage technology while focusing on the people.
During a 2024 interview (NBCUniversal News Group+1) he said, “Every day, 900,000 people go to our branches. They are much more advice centers than processing, operational centers.”
He acknowledged that while digital banking will continue to grow, many customers “want to have the backup for the branch. Customers like to visit their money.”
Dimon’s point is very straightforward, reinforcing the point that customers value human interactions with real people for complex issues.
In the end, the lesson of the ATM is not that machines replace people, but that they redefine where human value lies.
As automation reshapes contact centers, the organizations that succeed will be those that use technology to elevate human judgment because when it matters most, customers still want a human to say, “Yes, this is understood, and I’ve got you.”