The conversation about the future of contact centres often starts with technology. AI, automation, agent assist, routing, analytics and self-service dominate much of the current debate. But in our latest Contact Centre Network discussion, we brought the focus back to where it arguably belongs: the agent.
I was joined by Martin Teasdale, founder of Get Out of Wrap and the Team Leader Community, and Xander Freeman, Director at Call Centre Helper, to explore what contact centre agents really need in 2026 and how organisations can nurture them properly. The session was shaped by Martin’s recent Voice of the Agent research with Calabrio, which surveyed over 500 agents across the UK, Europe and the US. What made the discussion so valuable was that it wasn’t based on leadership assumptions alone. It was grounded in the direct voice of the people doing the job.
The central theme was clear: agents are not becoming less important in the age of AI. They are becoming more important. But the role they are being asked to perform is changing, and unless organisations rethink how they recruit, train, support, measure and lead agents, they risk creating a workforce that is technically enabled but emotionally exhausted.
From Transaction Handlers to Emotional Problem Solvers
One of the biggest shifts discussed was the move from agents handling largely transactional queries to managing more complex, emotional and high-stakes conversations. As AI and self-service take on simpler tasks, the contacts that reach a human are increasingly those where the customer is confused, frustrated, vulnerable, or unable to resolve their issue through digital channels.
Martin challenged leaders to recognise the scale of that shift before agents even enter the building. Recruitment and induction processes need to reflect the reality of the modern role, not the version of the job that existed ten or fifteen years ago. As he explained, “we need to be very, very clear what that’s going to look like and what it’s going to be like,” including both the rewarding parts of the role and the fact that “it’s going to be a challenge.”
That point matters because many organisations still present contact centre work as if it is primarily about being friendly, following process and handling volume. Those things still matter, but they are no longer enough. The future agent must handle nuance, emotion, judgement and complexity. That means recruitment should look more closely at resilience, curiosity, communication style, emotional regulation and problem-solving. It also means onboarding should prepare people for the emotional reality of the work, rather than simply teaching systems, scripts and compliance.
Xander reinforced this by pointing out that the traditional rhythm of the agent day is disappearing. In the past, agents may have had a mixture of simple and difficult calls. The easier calls gave them breathing space. As he put it, those “easy calls” were the ones where an agent could “reset” mentally. But those contacts are increasingly being automated away. His warning was clear: “You can’t just assume anymore that agents will go through the exact same things as you did back in the day.” That has big implications for workforce design. If organisations remove the simple work, they must deliberately build in recovery time, support mechanisms and better coaching, otherwise they risk creating a role where every interaction carries emotional weight.
Change Must Be Done With Agents, Not To Them
The panel also explored how organisations should support existing agents who are already in role and now being asked to work differently. This is a critical point. It is one thing to recruit future agents with new skills in mind. It is another to support experienced agents who have built confidence in one way of working and are now being asked to adapt to new tools, new workflows and new customer expectations.
Martin’s advice was practical and grounded in change management. He argued that agents need to be involved early and meaningfully, not simply informed once decisions have already been made. “These aren’t things that are going to be done to frontline staff and agents,” he said. “They will be involved in what this might look like.” That distinction is important. When change is imposed, it can create resistance, fear and disengagement. When agents help shape the change, they are more likely to understand it, trust it and improve it.
This is especially important because agents often know where processes break down better than anyone else. They see the real customer friction. They know where systems slow them down. They understand the difference between a policy that looks good on paper and one that fails under pressure. If organisations are serious about redesigning the future agent experience, they need to ask agents what they need from the business if transactional work is removed and more complex work remains.
Xander added that the organisations doing this well are those that are becoming more proactive in how they manage change. He argued that contact centres can no longer afford to wait until problems emerge before responding. “It’s not good enough anymore to just be as reactive as we have been in the past,” he said. That applies to technology, knowledge management, coaching and role design. If leaders know that AI and automation will change the shape of work, they need to prepare people before the change lands, not after it has already disrupted confidence.
The Empathy Paradox: Are Agents and Leaders Talking About the Same Thing?
One of the most interesting findings from Martin’s research was what we described during the session as the empathy paradox. Agents see empathy as one of their strongest skills. Yet leaders often identify empathy as one of the biggest gaps in their teams. That disconnect raises an important question: are agents and leaders even using the same definition of empathy?
Martin suggested that the gap may be less about whether agents care and more about how empathy is defined, observed and measured. “What I consider empathy versus what the leaders might consider empathy can be very, very different,” he explained. For an agent dealing with complex and emotionally charged calls all day, empathy may mean staying calm, listening, acknowledging the customer and getting them to a resolution. For a leader reviewing a call, empathy may be judged against tone, specific phrases, quality criteria or an idealised version of how the call should have sounded.
This is where expectations become important. Leaders who are not handling customer conversations every day may underestimate the emotional effort required to show empathy repeatedly under pressure. Martin raised the question of whether organisations are clear enough about “what empathy-driven, empathetic responses from our team look like” while still maintaining “a real understanding of real life.” That phrase matters. Empathy cannot be treated as an unlimited resource. It needs to be supported, replenished and coached.
Xander added another useful lens by describing empathy and efficiency as sitting on a spectrum. Contact centres often say they want empathy, but their metrics, dashboards and workforce pressures may still reward speed. When leaders are busy, under pressure or removed from the frontline, their view of empathy can shift. As Xander put it, leaders can “lose that perspective” of what good empathy looked like when they were closer to the agent role. This is why calibration, coaching and quality frameworks need to evolve. If empathy matters, organisations must be clear about what it looks like in practice and make sure operational measures do not undermine it.
Empathy Exhaustion and the Emotional Load of the Role
The phrase empathy exhaustion became one of the most important themes in the session. It captures something many contact centre agents experience but that organisations do not always name properly. Agents are expected to listen, reassure, absorb frustration, respond professionally and remain emotionally available, often across back-to-back interactions. When the simpler contacts disappear and the emotionally difficult ones increase, that load intensifies.
I asked whether contact centres are redesigning performance metrics quickly enough to reflect this new reality. Xander’s answer was balanced. He made the point that traditional metrics did not come from nowhere. Leaders use them because they need some way to manage demand, performance and service levels. However, he argued that the challenge is now about context and communication. “It’s less around metric A versus metric B,” he said, “it’s more around how we actually communicate those metrics to the people that have huge influence on the contact centre.”
That is a useful distinction. Metrics themselves are not always the enemy. The issue is when they are applied rigidly to a role that has changed. Average Handling Time, for example, may still be useful for planning, but it becomes problematic when used as a blunt instrument to judge individual agent performance in complex conversations. If an agent is supporting a vulnerable customer, recovering a failed journey, or resolving a multi-layered problem, speed alone is a poor measure of value.
Martin took the conversation further by asking whether the industry needs to look beyond coaching and wellbeing interventions and consider the fundamental design of the role. If the job now requires more emotional labour, should organisations review pay, working patterns, break structures and shift design? “How long can somebody maintain contact dealing with customers where the problems are more and more complex?” he asked. “Surely it cannot be the same as when fifty or sixty percent of their customer contact was more transactional.” That is the kind of question the industry needs to face honestly. If the work has changed, the operating model must change with it.
Burnout, Breaks and the Role of the Team Leader
Burnout was a major theme in the research and the discussion. Martin explained that when agents were asked why they think others leave the contact centre, burnout came through as the number one reason. Yet when agents were asked what makes them happy at work and what keeps them engaged, the most influential factor was often their direct manager or team leader.
This creates a powerful conclusion: team leaders are one of the most important protective factors in the agent experience. They are the people who notice when someone is struggling, create a sense of belonging, translate business pressure into practical support and make agents feel seen. But too often, the very activities that enable this support are the first things cancelled when service levels come under pressure.
Martin was clear on this point. “What are the first things that get thrown out the window?” he asked. “They are coaching sessions. They are one-to-ones. They are the times that individual agents get face time with their team leaders.” That is a dangerous pattern. Those conversations are not administrative extras. They are where agents process difficult interactions, build confidence, talk about development and feel valued. Removing them may help the queue in the short term, but it damages resilience and retention in the long term.
Xander echoed this by highlighting the role of team leaders in protecting agents from the wider pressure of metrics and business expectations. He made the point that no agent starts their day wanting to do a bad job. Performance issues often emerge when people lack support, clarity or the right conditions to succeed. This is why team leaders must have both the time and capability to lead properly. If organisations want better agent experiences, they cannot keep treating coaching as optional.
Have We Used AI to Help Agents, or Just to Drive Efficiency?
The discussion then moved into AI and whether organisations have used it in the right way. There has been a strong industry focus on efficiency: reducing wrap time, automating summaries, improving routing and speeding up access to knowledge. These are useful improvements, but the panel questioned whether they go far enough.
I asked whether we have taken the easy wins from AI while leaving the hardest problems for agents to absorb. Xander’s answer was candid: “If you’re asking me, have we done today? I’ll say honestly, no.” He acknowledged there are outliers doing great work, but the broader industry still has more to learn. He also made a valuable point about learning from the “first losers”, organisations that have tried to implement AI, failed, and had to row back. Those stories are just as important as success stories because they reveal what happens when the technology is implemented without enough thought, training or process redesign.
Martin added a more cautious perspective. He suggested that not fully removing all transactional work may not be a bad thing if organisations are not yet ready for the consequences. If businesses automate too quickly without redesigning support, coaching, scheduling and role expectations, the change could shock the system. In his words, moving too quickly “could be the wrong thing to do” if the wider operating model is not ready.
This is an important conclusion. AI should not simply strip out workload and leave humans with the most difficult work without redesigning the environment around them. The question is not just what AI can automate. It is what kind of human work remains afterwards, and whether the organisation has prepared people properly for it.
AI Confidence Versus AI Fear
Martin’s research also highlighted a significant AI knowledge gap. Many agents do not clearly understand where AI is already being used, while more than half worry about its impact on their jobs. That creates a climate of uncertainty. The issue is not always resistance to AI. Often, it is a lack of clarity.
Martin argued that the starting point is communication. “They don’t know where it is being used at the moment,” he said. That is not a difficult problem for organisations to address, but it is often overlooked. Leaders need to explain where AI is being used, what it is doing, what it is not doing, and how it affects the agent role. Without that clarity, agents are left to fill in the gaps themselves, and those gaps are often filled with fear.
Xander suggested a practical way to build confidence: involve agents in pilots and let peer advocacy do some of the work. “Pilot this stuff with a small group of agents and word will spread,” he said. That is far more powerful than a senior leader simply announcing a new tool and expecting adoption. When agents can see colleagues using AI successfully and honestly discussing the benefits and limitations, confidence grows more naturally.
The panel also explored how AI could support coaching. Xander highlighted the growth of virtual training environments where agents can practise difficult conversations with AI before facing them live. This could be a valuable way to build confidence in high-pressure scenarios without risking customer outcomes. But both speakers were clear that AI should support human development, not replace the human relationships that hold teams together.
Will AI Replace the Team Leader? No , But It Will Change the Role
One of the more interesting moments came when we discussed whether agents might prefer coaching feedback from AI because it can feel less judgemental. This is already happening in some environments, where real-time tools and AI-generated feedback give agents immediate, impartial insight into performance.
Martin responded by reflecting on the way we are all experimenting with AI in our own lives. We may ask it for advice, structure or ideas, but most people do not follow it blindly. They apply human judgement. He suggested the same will be true in contact centres. AI may change the nature of team leadership, but it will not remove the need for team leaders.
His description was particularly strong: team leaders are “the people that create the esprit de corps of the team.” They are the ones who hold people together, create connection, build morale and notice the things a system cannot fully understand. He described them as the “NCOs of the brigade” , the people closest to the frontline who translate intent into action and keep the team functioning.
That is a useful way to think about the future. AI may take away some administrative burden. It may help with coaching preparation, summaries, quality insight and knowledge access. But the relational work of leadership , trust, belonging, encouragement, challenge and care , remains deeply human.
Technology, Neurodiversity and Cognitive Load
We also explored a topic that does not get enough airtime: the impact of increasingly complex technology environments on neurodiverse agents and those experiencing cognitive overload. Many contact centres are adding more dashboards, prompts, alerts, colours, pop-ups and real-time guidance. Leaders may see this as support, but for some agents it can become another layer of distraction.
Xander made the point that technology can be incredibly enabling when designed well. Speech-to-text improvements, for example, can make work more accessible for agents who are hard of hearing. But the idea of a “single pane of glass” can quickly become overwhelming if that pane is full of competing prompts and visual noise. His view was that organisations need “human guardrails” as well as technical ones.
That matters because personalisation should not only apply to customers. It should also apply to agents. Some people will benefit from real-time nudges and visual prompts. Others may need simpler interfaces, fewer distractions, different pacing or alternative ways of receiving support. If we want inclusive contact centres, we need to design technology around different cognitive needs, not assume one interface works for everyone.
I also raised the risk of routing the most complex contacts to the most capable agents. It sounds sensible operationally, but it may create another problem: the best agents become overloaded because they carry the emotional weight of the hardest work. Martin connected this to a familiar operational challenge. In the past, agents sometimes resisted upskilling because it meant receiving harder work without enough benefit or recognition. The same risk exists now. If we route complex work to “super agents”, we must ensure they feel valued, supported and rewarded , not simply used.
The Human Advantage in an AI World
The session closed with the big question: if AI makes agents more indispensable rather than less, what must organisations do to protect and nurture the human advantage?
Xander’s answer focused on leverage. He argued that contact centres have more strategic value than ever because they hold the data the wider business needs to make AI work. “I can’t remember a time where contact centres had more leverage,” he said. “Everyone is under pressure to implement AI… and when they need that data, they get it from the contact centre.” This is a powerful reframing. The contact centre should not see itself as a cost centre waiting to be automated. It should see itself as the source of customer intelligence and human expertise that makes better automation possible.
He also made a useful comparison: “Accountants didn’t go extinct the minute we invented calculators.” Instead, the accountants who learned to use calculators became more effective. The same applies to agents. Those who learn how to work with AI, interpret insight, and use technology to improve human conversations will become more valuable, not less.
Martin brought the conversation back to the customer. At its heart, he said, the contact centre exists to respond to customer needs and help people. Customers still want to speak to human beings, especially when the issue is complex, emotional or important. That means skills like empathy, rapport, reassurance, humour and relationship-building will matter more, not less. “Customers still want to talk to human beings,” he said. The opportunity is to create “more competent, better trained, better rewarded, more skilled professionals” who can deliver the human contact customers still value.
Final Thoughts: Better CX Depends on Better Agent Experience
What stood out most from the discussion was that the future of the agent role is not simply a technology story. It is a human sustainability story. If AI removes the easy work, then agents need better support for the harder work. If organisations expect more emotional intelligence, then they must protect agents from empathy exhaustion. If leaders want agents to trust AI, they need to communicate clearly and involve them early. If businesses want excellent customer experience, they must stop treating agent experience as secondary.
The role of the agent is becoming more skilled, more emotionally demanding and more central to the success of the organisation. That should change how we recruit, how we train, how we schedule, how we measure, how we coach and how we lead.
The most successful contact centres in 2026 will not be the ones that simply automate the most. They will be the ones that use technology to create more space for human value. They will listen to agents, protect their energy, support their growth and design work that recognises the emotional reality of the role.
Because the future of customer experience will not be powered by AI alone. It will be powered by confident, capable and well-supported people who know how to use AI without losing the human connection customers still need.
