Contact center AI adoption is accelerating, but UJET’s recent customer visits suggest speed alone is a weak measure of progress. This emerged as a central theme during CX Today’s interview with Kristin King, Chief Customer Officer at UJET, and Tena Curic, Senior Customer Success Manager at UJET.
The conversation examined what UJET learned after spending time inside customer contact centers, sitting with agents and supervisors, observing real workflows, and asking what teams actually need from AI.
It also challenged a familiar pressure point in today’s CX market: the push to “do AI” quickly, even when organizations have not clearly defined the outcome they want.
For King, many AI projects start to lose their way when leaders feel pressure to move before they have defined the result. In several customer conversations, she saw teams being pushed to adopt AI without clear goals, clean data, or internal alignment. King framed the pressure as a leadership challenge, not simply a technology decision:
“They’re getting pressure from the board, they’re getting pressure from competitors, they’re getting pressure from the market, and they want them to add AI.”
Teams are getting stuck when the data isn’t clean, the data is not connected, they don’t have a clear understanding of how AI should be used.
Understanding this gap is vital now that AI adoption is now a boardroom expectation, a competitive marker, and a budget conversation. Yet in the contact center, rushing into AI without clear operational intent can add complexity rather than remove it.
Why Contact Center AI Adoption Is Under Pressure
CX leaders are operating in a noisy market. Boards want visible AI progress, competitors are experimenting, vendors are pushing new capabilities, and customers expect faster, more personalized support.
At the same time, contact centers are being asked to improve quality, reduce cost-to-serve, support leaner teams, and modernize operations. AI appears to offer a route through that tension. However, UJET’s field work suggests the starting point matters.
Leaders need to understand which business outcome they want AI to improve before they decide which tool to deploy. Otherwise, the contact center risks buying technology before it understands the friction it is trying to remove.
This is exactly where AI programs can become dangerous. A project may look ambitious and move quickly, while still failing to solve a defined problem for customers, agents, or supervisors.
The Data Foundation Problem
One of the most common barriers to responsible AI adoption is data readiness. Contact centers often sit on large volumes of customer interaction data. However, volume rarely translates into data that is clean, connected, or ready to support AI-led decisions.
Customer conversations may live in one place. CRM data may live somewhere else. Quality data, workforce data, tickets, survey responses, and escalation histories may sit across separate systems.
When AI is layered on top of that fragmentation, it can expose the problem rather than solve it. King warned that weak foundations can turn AI into another source of friction:
“AI can actually cause more friction and a poor customer experience. But conversely, when it’s well thought out, when you know the outcome that you’re trying to drive to, the goal that you’re trying to drive to, you can then create a best-in-class experience for your customer and for your agent at the same time.”
This operational reality is critical. AI can improve experiences when it has the right context, the right workflow logic, and the right operational goal.
Without that foundation, it can route poorly, answer incompletely, summarize inaccurately, or force agents to correct mistakes after the fact. At that point, the technology meant to reduce friction becomes another source of it.
Start With The Floor, Not The Tool
Curic’s onsite observations add an important operational layer to this leadership challenge.
Before leaders decide which AI tools to deploy, they need to understand what actually slows teams down. Achieving this requires sitting with agents, watching how they work, and noticing the effort that dashboards often miss.
Agents may spend time moving between customer information, billing tools, ticketing systems, and internal notes while still trying to keep the customer conversation natural. Those moments are easy to overlook from a distance. On the floor, they become obvious. Curic argued that leaders should begin with the work agents experience every day:
“Before asking what tool should I buy, try asking what is actually slowing our people down every day, because the agents know exactly where the friction is.”
Reframing the problem this way changes the AI conversation. Instead of starting with a product category, leaders start with the work. They look at repeated lookups, manual notes, system-hopping, status checks, routing friction, escalation gaps, and the places where agents create workarounds.
Those details may seem small in isolation. Across thousands of interactions, they can reveal where AI may deliver fast and safe value. They also help leaders avoid automating the wrong process.
Agent Buy-In Is An Adoption Requirement
AI projects can struggle when agents feel they are being acted upon rather than included. This risk is particularly high in contact centers, where new tools can change workflows, performance expectations, and customer interactions quickly.
If agents do not understand why AI is being introduced, or if they do not trust the output, they may work around it. The result is a familiar problem: leaders think they have deployed new capability, while frontline teams quietly continue with old habits.
UJET’s customer visits suggest agents are not rejecting AI outright. They are asking for it to be aimed at the right work. They want help with repetitive tasks, context gathering, summaries, routing, and routine checks.
They are more cautious when AI touches sensitive interactions, emotional conversations, or moments where customer trust is on the line.
Move Fast, But Keep Direction
There is still a strong case for speed. CX leaders cannot spend years debating AI strategy while customer expectations change and competitors experiment, but speed needs direction. King put the warning simply:
“Don’t confuse speed with progress. It’s important to move fast on AI, but not to move in the wrong direction.”
Ultimately, this serves as the most useful takeaway for CX leaders under pressure this year. Progress is not the number of AI tools deployed. It is not the number of workflows automated. It is not a pilot announcement.
Progress shows up when customers experience less effort, agents gain useful support, supervisors see clearer patterns, and leaders can connect AI investment to measurable outcomes.
Success demands a practical sequence: define the problem, understand the workflow, validate the data, involve the people doing the work, and deploy AI where it can improve the experience.
Partnership Beats Prescriptive Selling
The interview also surfaced a broader point about vendor relationships. Customers do not always need another software to pitch. Often, they need help making sense of where to begin.
This holds especially true in AI, where the market is noisy, and the risks are not always obvious at the start. King and Curic described the value of being physically present with customers, sitting inside the operational reality before proposing an answer.
This physical presence is essential because contact center problems rarely appear neatly packaged. They show up in repeated friction, informal workarounds, agent habits, supervisor concerns, and customer frustration that may not appear in a standard quarterly review.
For UJET, the lesson is that partnership begins before the solution is obvious. It begins in the messy part of the work, where leaders, agents, and technology teams are still trying to understand what is really happening.
Move Fast, But Know Where You’re Going
The next stage of contact center AI adoption will likely reward teams that move with discipline and that requires moving with a sharper understanding of the problem.
Leaders should start by asking what outcome they want AI to improve. Then they should test whether the data, process, and frontline confidence exist to support that outcome. Only then does the technology decision become meaningful.
The danger for contact centers is acting quickly, spending heavily, and leaving the core friction untouched. AI can help contact centers move faster, serve better, and reduce unnecessary work. Yet those gains depend on knowing where the friction lives and why it exists.
