AI customer service is moving quickly, but customers are not asking brands to replace empathy with efficiency.
That is the tension at the heart of ServiceNow’s latest CX research. According to The CX Shift, 53 percent of customers expect AI to improve speed and efficiency, while 50 percent still cite a lack of empathy as their top frustration.
For CX leaders, that creates a difficult question. Can AI make service faster without making it feel colder, more transactional, or less accountable?
The answer depends less on the AI itself and more on what surrounds it. Mark Ashton, VP of Solution Consulting, CRM at ServiceNow, framed the challenge as a shift in expectations:
“Everything we do is judged on this kind of speed and empathy. We want speed, but we still love talking to people.”
That matters because the customer service conversation has changed. Customers increasingly compare every service interaction with the fastest and most intuitive experiences they receive anywhere, from digital retail to ride-hailing apps.
But when a customer is stressed, vulnerable, confused, or dealing with a complex issue, speed alone rarely solves the problem.
Why AI Customer Service Needs More Than Speed
The promise of AI in customer service is easy to understand. It can summarize conversations, retrieve information, automate repetitive tasks, and help customers resolve simple queries without waiting in a queue.
ServiceNow’s research suggests customers are already open to that shift. Many want faster answers and more efficient journeys, particularly for routine service tasks.
Yet the same research shows that speed does not remove the need for empathy. Customers still want to feel heard, understood, and supported when the stakes are higher.
Ashton argued that organizations need to be deliberate about where AI belongs and where human judgment remains essential:
“AI is good at repetitive, transactional, information-heavy interactions. Human agents are really good at empathy, complexity, listening, emotional stress, and triggers.”
That distinction should shape how enterprises design AI-enabled service. AI can help with order tracking, account updates, summaries, and routine requests, while humans remain central to complex, sensitive, or high-value moments.
The risk is that businesses chase efficiency too aggressively. If the goal becomes cutting handling time or reducing headcount without rethinking the experience, AI may improve internal metrics while weakening customer trust.
The Empathy Gap Is Also an Operations Problem
It would be easy to treat empathy as a soft skill issue. In reality, it is also a systems issue.
ServiceNow’s report found that service reps spend only 45 percent of their time addressing customer issues. It also found that 80 percent have to log into three to five systems to resolve a single customer problem.
That fragmentation affects the agent and the customer. When agents have to search across systems, repeat questions, and manually connect information, customers feel the friction.
Ashton described this as a context problem. If an agent can see whether someone has called three times, is making a high-value purchase, or is dealing with a sensitive life event, they can respond differently.
Without that context, even a well-intentioned agent may sound detached. Ashton pointed to the operational root of the problem:
“A great word here is context. If we can give the agent that context, then it is a lot easier.”
That is where AI can help service feel more human. It can give agents a clearer picture of the customer, surface relevant history, summarize previous interactions, and reduce the administrative burden around the conversation.
However, that only works if the organization has connected the right data and workflows behind the scenes.
When Humans Become The Middleware
The wider problem is that many enterprises still rely on people to hold fragmented service journeys together.
In CX Today’s accompanying video interview with Ashton, the discussion focused on why customer experience has outgrown traditional CRM and why enterprises now need to connect the front, middle, and back office. That same issue shows up clearly in the service rep experience. Ashton put it plainly:
“The human is acting as the middleware, connecting the dots between these systems, not the IT or the application itself.”
That phrase captures the challenge for many service organizations. Agents are not only handling customer emotions. They are also bridging gaps between systems, departments, processes, and data sources.
AI can reduce that burden, but only if it connects to the right enterprise foundation. Otherwise, it risks becoming another layer for agents to manage.
For CX leaders, this means AI strategy cannot sit apart from workflow strategy. Service leaders need to understand where work gets stuck, which systems agents rely on, and where customer context disappears.
They also need to consider how AI changes the rhythm of the agent’s day. Ashton shared an example of a utility provider that used AI to summarize calls and reduce wrap-up work during a natural disaster.
The technology helped productivity, but it also increased talk time. Agents had fewer natural pauses, and the business had to reintroduce space into the process so teams could maintain empathy during emotionally demanding calls.
That example matters because it shows how AI can improve one metric while creating pressure somewhere else.
AI Should Add To Humans, Not Simply Replace Them
The strongest AI service strategies are unlikely to remove people from the experience altogether. Instead, they will remove the work that prevents people from doing their best work.
That may include summarizing calls, updating CRM records, retrieving case information, detecting patterns, and guiding agents through next-best actions. Ashton said this is already shaping executive discussions:
“In nearly every executive meeting I have, we are still talking about how we help humans all the time.”
That point is important. Many AI conversations start with automation, but enterprise service leaders also need to ask how AI changes the value of human work.
If AI handles repetitive tasks, service teams can spend more time on complex cases, vulnerable customers, loyalty risks, and commercial opportunities.
But leaders need to resist a narrow productivity calculation. If a specific AI use case reduces 10 percent of inbound demand, the strategic question is not only whether the business needs fewer people.
It is also whether those people can be redeployed to deliver better service, handle more valuable interactions, or improve customer retention.
That is where the service function starts to shift from a cost center toward a customer value engine.
From Customer Record To Customer Context
Traditional CRM helped organizations record what happened. The next phase of AI-enabled CX depends on whether businesses can use those records, signals, and workflows to support action in the moment.
Ashton described old-school CRM as “writing a newspaper of what happened” rather than using signals to understand what should happen next.
That gap is increasingly important in a connected economy. Many products and services now generate ongoing usage signals, from subscription software to connected devices.
If a customer suddenly stops using a service, changes behavior, or shows signs of frustration, AI can help detect the signal and trigger a proactive response.
That is where service becomes more personal without relying only on the agent’s memory or manual effort. The system can surface relevant context before the customer has to repeat themselves.
For customers, that can make the interaction feel more human. For agents, it can make empathy easier to deliver.
What CX Leaders Should Do Next
AI can make customer service faster, but speed should not become the only measure of progress.
The more useful goal is resolution with context. Customers want simple service, but they also want brands to recognize the situation, understand the history, and respond with the right level of human care.
That requires connected data, integrated workflows, and a clear view of where AI should assist, where it should automate, and where people should lead.
It also requires a staged approach. In our next article we explore autonomous CX, and why automation can reduce complexity only when enterprises have the right foundations, governance, and trust in place.
For now, the priority is simpler. CX leaders should look at where agents are forced to act as middleware, where customers have to repeat themselves, and where AI can remove friction without removing empathy.
