Contact center AI is often framed as a replacement story. UJET’s recent visits to customer contact centers suggest the sharper opportunity is much more practical: remove the work that slows agents down.
That was the central thread running through CX Today’s recent conversation with Kristin King, Chief Customer Officer at UJET, and Tena Curic, Senior Customer Success Manager at UJET.
The conversation brought together King’s customer strategy perspective and Curic’s onsite observations from recent contact center visits.
It explored what UJET learned after spending time inside customer contact centers, sitting with agents and supervisors, listening to calls, and observing how work really happens between the clean lines of a dashboard. Curic described the reality from the contact center floor:
“What stood out to me is how much of the job actually happens in between those obvious moments.”
Understanding this gap is important because the industry often measures the result of an interaction, then misses the effort it took to get there.
A dashboard can show that a call ended successfully. It cannot always show the agent moving between customer records, billing systems, internal tickets, and knowledge sources while trying to sound calm, natural, and present.
For CX leaders, that hidden effort is becoming one of the most important places to look.
UJET’s Contact Center AI Lesson: Start With The Work Agents Actually Do
Curic’s onsite observations were striking because they were not built around abstract AI strategy. They were grounded in what agents experience every day.
She described sitting beside agents who were calm, knowledgeable, and empathetic, while still spending large parts of each interaction trying to assemble context from disconnected systems. That is where AI can help, if leaders point it at the right problem.
Curic emphasized that agents were not resisting technology. They wanted help with the work that added time and friction without requiring human judgment. Across the visits, Curic heard a consistent message from frontline teams:
“Help me move faster through the noise so that I can focus on the customer.”
It’s a refreshing reality check for a market conversation that often jumps straight to automation targets. Agents do not need AI to take over every conversation. They need it to reduce repetitive lookups, status checks, note summaries, routing, and the manual steps that make good service harder than it should be.
Those tasks may look small in isolation. Across hundreds or thousands of contacts, they become a tax on agent focus, customer patience, and service quality.
Workarounds Are Warning Signs
One of the most valuable observations from the interview was Curic’s point about workarounds. When a tool does not quite fit a process, agents may not raise a formal issue. They often build manual steps around the gap and keep moving.
This workaround is familiar to anyone who has worked in a contact center but it is also easy to miss from the outside.
Workarounds can look like resilience, and in many ways they are. Yet they can also signal where technology is creating extra work, where processes have drifted, or where agents have learned to compensate for systems that do not support the way customers actually behave.
For UJET, the lesson from sitting on the floor was that these moments deserve attention. They reveal where AI can remove friction without weakening the customer relationship. They also show where automation could make the wrong problem worse if leaders do not understand the work first.
Remove The Right Work
King framed the issue as a question of control, context, and judgment. The market narrative often suggests that AI progress means removing agents from more interactions. UJET’s conversations with customer teams pointed in a different direction.
King argued that the strongest use of AI is to remove lower-value work while keeping agents available for the moments where human judgment shapes the outcome::
“The goal is to remove the right work and not just removing work.”
This is where context becomes everything. A customer with a simple status question may prefer a fast automated answer. A customer who is frustrated, confused, emotional, or stuck in a repeat issue may need a human who can read the situation and adjust.
The task for CX leaders is to decide which work drains agent capacity, and which work carries relationship value. AI can automate repetitive tasks, it can surface context, it can summarize notes. It can even help with routing and handoffs. But when timing, empathy, escalation, and judgment matter, customers still expect the experience to feel human, even if technology is supporting it behind the scenes.
Customers Notice Outcomes, Not The Tool
Curic also raised a useful point about personalization. The best AI experiences, she suggested, are not remembered as AI experiences. They are remembered as good customer interactions.
She shared the example of a retail underwear brand using AI to personalize service based on customer history, preferences, and sizing information. Customers were leaving positive reviews. Yet they were not praising the technology. They were saying they felt seen, understood, and treated as individuals.
This represents a more mature way to evaluate AI in CX. Ask whether the interaction felt relevant, accurate, and easy. If AI helps the brand understand a customer faster, the outcome may feel more personal. If AI blocks the customer from the help they need, it becomes another layer of friction.
For contact centers, that should change the success criteria. Deflection alone is not enough. Lower chat volume is not enough. The better test is whether customers get the right outcome with less effort and more confidence.
Efficiency Cannot Break Trust
The pressure on contact centers is real. Teams are leaner, budgets are tighter, and leaders are being asked to do more with less while improving service quality. That creates a natural opening for AI but it also creates risk.
Curic noted that teams are open to efficiency when it targets work nobody wants to do anyway, such as basic information lookup, routing, and note summarization. But she also said teams were much more protective of the moments that shape how customers feel, highlighting the line teams were not willing to cross:
“Teams do want efficiency, but not at the expense of quality or trust.”
This should serve as the strategic anchor for CX leaders to keep coming back to. AI should give agents time back, it should reduce context hunting and it should help supervisors see where process friction lives.
It should not make customers feel trapped, misunderstood, or pushed away from the help they need.
Before You Automate, Watch The Work
The immediate takeaway from UJET’s floor visits is simple, and perhaps slightly uncomfortable. Before asking which AI tool to buy, leaders should ask what slows their people down every day.
That question changes the shape of the project. It forces teams to look at the operational reality behind the metrics. It invites agents and supervisors into the conversation earlier. It helps leaders target AI at the work that creates friction, rather than the work that looks easiest to remove.
It also creates a stronger bridge between customer experience and employee experience. Agents who spend less time fighting systems can spend more time helping customers. Customers who receive better context and faster resolution feel the difference, even if they never know which technology made it possible.
Ultimately, this reveals the quieter, more useful promise of contact center AI. It may not begin with replacing people. It may begin with finally removing the clutter that kept them from doing their best work.
