I recently witnessed a leadership meeting where an agenda item grew into a major, heated discussion. Should a common customer request be handled entirely by AI, or does a person need to stay in the loop?
It was supposed to be a quick decision. Get it wrong one way, and you frustrate customers with unnecessary friction. Get it wrong the other way, and you remove human oversight from an interaction that occasionally surfaces something serious.
Midway through, one of their managers said something that stuck with me: "This used to be an IT ticket. Now it's a values decision."
That's a good way to put it. For years, contact center leaders have been told that AI will change their operations. That's true, but it misses the bigger change: AI is changing the job of leading a contact center.
Decisions that once seemed relatively straightforward are becoming interconnected. What should be automated? Who is accountable when AI gets something wrong? What skills will employees need? How do you plan when both the volume and nature of the work are changing?
These aren't primarily technology questions. They're leadership questions.
ICMI's State of the Contact Center in 2026 illustrates the range of expectations. Fifty-eight percent of leaders predict AI integration will result in moderate to substantial staffing reductions over the next two to three years. Another 39% anticipate staffing levels to remain similar while agent responsibilities shift toward higher-value work.
Both outcomes are plausible. AI will undoubtedly automate work. But the work that remains is becoming more complex, consequential, and human. At the same time, leaders are being asked to make decisions about technologies that are advancing faster than the policies, skills, and management practices surrounding them.
The challenge isn't simply adopting AI. It's leading through what AI changes.
Start With the Customer, Not the Technology
The temptation with every major technology wave is to begin with the technology.
Where can we use AI? What can we automate? How quickly can we deploy it? How much can we save?
These are reasonable questions, but start instead with what customers are trying to accomplish. Which interactions, or steps in interactions, are genuinely better automated? Where can AI help employees make better decisions? Where is human judgment essential? And what problems shouldn't exist in the first place?
AI makes it possible to automate enormous amounts of activity. Leadership determines whether that activity should exist at all. The best leaders aren't looking for places to deploy AI. They're redesigning work around the outcomes customers and the organization need.
Redefine the Human Role
Few questions generate more attention than what AI will do to contact center jobs. But "Will AI eliminate jobs?" is too simplistic. A more useful question is: What work should humans do when AI can do much of what humans used to do?
ICMI's research suggests how leaders expect the answer to evolve. As AI takes on more routine work, respondents see agent roles shifting toward higher-value responsibilities. Employees will spend more time solving problems, handling exceptions, applying judgment, and providing personalized assistance. They will also need much greater comfort working alongside AI-assisted tools.
The human role will become more consequential. Consider a healthcare provider that sees AI absorb many routine scheduling and billing inquiries. What’s left for agents? The issues no algorithm resolved. A family member trying to navigate a loved one's care, patients caught between conflicting instructions from two providers, a billing dispute tangled in insurance rules. Customers may try more than once and receive different answers from the "same" system.
You have to spend time in a contact center to see this work firsthand. It's often the exception. The unusual circumstance. The customer who has already tried three things. The emotionally charged conversation. The decision requiring context that doesn't fit neatly into a predefined scenario.
These interactions require more from employees, not less. Leaders will need to rethink hiring profiles, development, coaching, empowerment, career paths, and even what good performance looks like. As the work changes, the way we lead our people and teams has to change with it.
Build Capability, Not Just AI Skills
Whenever a new technology arrives, training tends to focus on how to use it. That's necessary, but nowhere near sufficient for AI.
The larger need is organizational readiness. Agents need to know when to rely on AI, question it, or override it. Supervisors need to coach AI-assisted work. Quality teams need to evaluate outcomes produced jointly by humans and AI. Workforce planners need to understand how automation changes workload, while knowledge teams must ensure AI has accurate, current information.
Leaders don't have to become data scientists or AI engineers. But they need enough fluency to ask good questions about capabilities, limitations, risks, customer impact, and business value. It is essential to create an environment where people can keep learning as the technology evolves. AI readiness isn't a one-time training event; it’s an ongoing organizational capability.
Create Accountability Without Slowing Everything Down
As AI becomes more capable, another leadership question becomes unavoidable: Who is accountable?
When AI suggests a response that an employee reviews before sending, the answer may seem straightforward. But what happens when AI summarizes interactions, recommends next steps, predicts customer intent, routes work, makes decisions, or begins completing actions autonomously? The governance questions multiply quickly.
In one example, an airline's system automatically rebooked a family after a modest schedule change, creating a bizarre connection and much longer travel time even though seats remained on the original flight. During a lengthy call from one of the family members, an agent restored the itinerary and waived change fees. But there was no clear accountability for the underlying problem or authority to intervene quickly. (The company created an escalation path for regularly reviewing edge cases and refining the rules.)
ICMI's research is revealing here. The biggest concerns leaders identify around implementing agentic AI aren't primarily employee resistance. Customer resistance and the accuracy and integrity of AI data were each cited by 47% of respondents. Data privacy and security risks and integration with existing systems followed at 46%, while 45% identified a lack of proper AI governance and 41% cited a lack of internal expertise.
These are leadership issues as much as technical ones. Organizations need clarity around what AI can decide, what requires human review, how accuracy and outcomes are monitored, when employees should override recommendations, how problems are surfaced, and who owns corrective action when something goes wrong.
And there's another side to this. Governance can become so cumbersome that organizations become afraid to experiment. The answer isn't less accountability, it’s clearer accountability. The best leaders create guardrails that allow responsible experimentation while protecting customers, employees, and the organization. They know the difference between moving thoughtfully and simply moving slowly.
Lead Across Boundaries
Few of these decisions belong entirely to the contact center. Customer journeys cross functions, AI draws on data and knowledge from across the enterprise, and automation decisions affect customer experience, staffing, risk, and brand trust.
Contact center leaders increasingly have to influence all of them. The strongest leaders are building relationships across technology, operations, HR, finance, marketing, product, and other areas. They're bringing customer intelligence into decisions made elsewhere. They're identifying where policies, processes, products, and digital experiences create unnecessary demand. And they're helping the organization decide where technology can simplify customer journeys rather than simply automate pieces of them.
This broader role is consistent with another finding in ICMI's research. Seventy-four percent say their organizations treat the contact center as a top priority for the business, while 62% believe other departments see it as a customer experience hub. Forty-eight percent say it is viewed as a strategic business partner. AI makes that organizational role more important, not less.
The contact center sits at the intersection of customer needs, the organization’s processes, employee capabilities, data, and emerging technology. Few areas of the organization have a better vantage point for seeing how all of those pieces actually come together for customers. That gives contact center leaders the opportunity (and responsibility) to influence decisions far beyond their traditional boundaries.
Leadership Becomes the Differentiator
AI capabilities will continue to improve. Over time, access to the technology itself won’t be much of a differentiator. What organizations do with it will matter far more.
The defining skill for the next generation of contact center leaders won't be knowing more about AI than everyone else. It will be knowing how to lead in an environment in which AI is everywhere.
That means making thoughtful choices about what should be automated and where human judgment matters. It means preparing employees for work that is more complex and consequential. It means establishing accountability without suffocating innovation. And it means connecting decisions about technology, people, customer experience, and business strategy that many organizations have traditionally made separately.
AI is making contact centers more technologically sophisticated. But it's also making leadership more important.
The organizations that pull ahead won't necessarily be those that adopt AI fastest. They'll be the ones whose leaders are best prepared to turn its capabilities into better customer experiences, stronger employees, simpler operations, and better business results.
Read Brad's other articles in the series:
The contact center knows your customer best. Are you using it that way?
The Best Contact Centers Have Outgrown the Traditional Metrics. Here’s What They Measure Instead