AI is changing what’s possible in customer experience. But are we changing the way our businesses operate quickly enough to take advantage of it?
For years, we’ve been asking how to make customer service more efficient. AI is giving us some extraordinary answers. We can resolve more interactions without human involvement, give agents better information, automate routine tasks, and reduce the effort involved in supporting customers. All of which is progress. But I wonder whether we’re asking a question that’s becoming too small.
Take a customer whose flight has been canceled. We can use AI to answer their questions, explain their options, and help them rebook. That’s a meaningful improvement on spending an hour waiting to speak to someone. Or a bank can flag suspicious activity and immediately block a transaction from happening.
Take the travel example: could we have anticipated the disruption? Could we have offered alternatives automatically? Could we have updated their itinerary and kept them informed without requiring them to navigate the process themselves?
The second scenario is where things get interesting. Not necessarily because it requires more sophisticated AI, but because it requires the business to work differently. Customer service, operations, digital, data, and technology all need to come together around the customer. We call this orchestration, and that’s where I think the next generation of CX faces one of its greatest challenges.
We’re introducing technology that can work across the customer journey into organizations still largely structured around individual functions. Customer service owns the contact center. Digital owns the website. IT owns the systems. Product owns many of the decisions that determine why customers need support in the first place. Everyone owns a piece of the experience. Very few people own the whole thing.
It’s a tension I recently explored at the Execs in the Know Customer Response Summit, opening a discussion on The Next Gen CX Business Plan: Preparing for the Next 3–5 Years.
AI is exposing problems we’ve learned to work around
Anyone who has worked inside a large customer operation will recognize this.
A customer contacts support with a relatively straightforward problem. The agent needs to check one system for account information, another for an order, perhaps a third for the latest policy. They may need to contact another department or find someone who can resolve an unusual exception. Eventually, they piece together the information and help the customer.
From the outside, the interaction might look successful. The issue is resolved, the customer receives an answer, and the business moves on. But look at what actually happened. We relied on a person to compensate for disconnected systems, fragmented information, and processes that don’t quite work together.
We’ve become remarkably good at doing this. In many ways, it’s what has made large-scale customer operations possible. Now we’re asking AI to navigate those same environments, often with far less human involvement. And suddenly, the weaknesses in our operating models become much harder to ignore.
An AI assistant can only be as useful as the information it can access and the actions it’s authorized to take. If policies contradict each other, systems don’t communicate, or nobody has established who owns a particular decision, introducing AI won’t resolve those issues.
It may simply expose them more quickly.
This is something we’ve explored extensively through Transcom’s AI, but Make It Human series, where Jeff Mortlock, SVP, Solutions, and Cortney Jonas Burnos, VP Digital Solutions, examine what AI implementation looks like inside real customer operations.
One recurring theme is how easily the conversation outpaces the fundamentals. Organizations are exploring increasingly sophisticated AI applications while opportunities to improve knowledge management, coaching, training, and everyday workflows remain relatively untapped. It’s tempting to focus on the most ambitious use cases. But the less glamorous work often determines whether those ambitions are achievable.
Research from McKinsey reinforces this distinction. Its analysis of the AI productivity paradox points to organizational changes, including workflow redesign, clearer ownership, and revised decision-making responsibilities, as essential to turning AI investment into meaningful business performance.
For CX leaders, that suggests the technology roadmap and the operating model can no longer be treated as separate exercises.
Are we improving the experience, or just automating the problems?
Making customer service easier to access isn’t quite the same as improving the customer experience. Consider a business receiving thousands of contacts every month about confusing bills. An AI assistant could answer those questions quickly and accurately, perhaps resolving most of them without human intervention. That’s a valuable improvement. Customers spend less time waiting, and the business reduces the cost of handling those interactions. But the bills are still confusing.
We’ve made it easier to deal with the problem without doing anything about the problem itself. And because the contact center is now handling fewer interactions, we might even consider the initiative a success. This is where our traditional performance measures can fall short.
If we’re primarily measuring containment rates, handling times, and cost per contact, we’re likely to focus on making existing interactions more efficient. But if we’re interested in customer effort, trust, and the overall quality of the experience, we have to look further upstream.
Why are customers contacting us? Which problems keep recurring? What would it take to eliminate them? The answers often sit outside the contact center.
Perhaps the billing team needs to redesign a statement. Perhaps product needs to address a recurring defect. Perhaps digital needs to simplify a journey that customers consistently abandon. None of these improvements necessarily requires AI. But AI can help us identify patterns, understand customer frustration and make better-informed decisions about where to intervene.
The real opportunity is to connect those insights with the parts of the business that can act on them. Otherwise, we risk spending the next five years becoming exceptionally efficient at managing customer problems that shouldn’t exist.
There’s also a workforce dimension to consider.
Gartner reported in April 2026 that 85% of surveyed customer service leaders were expanding agents’ responsibilities as AI reduced contact volumes and changed the nature of frontline work. That makes sense. As straightforward interactions become easier to automate, the work left to people increasingly requires judgment, problem-solving and an understanding of the wider customer journey.
Yet many organizations are still managing and developing their teams around models designed for a different mix of work. We need to think carefully about the capabilities people will require, how we support them and whether our measures of performance reflect the value they’re being asked to deliver.
In our AI, but Make It Human conversations, we return frequently to the importance of AI literacy. Not simply teaching people how to use a tool, but helping them understand its limitations, question its recommendations, and exercise judgment about when human intervention is needed. Those skills become more important, not less, as AI takes on a greater share of customer interactions.
A different starting point for the next five years
So where does this leave CX leaders planning for 2030? I don’t think we need another reminder to embrace AI. Most organizations are already investing in it, experimenting with it, or trying to understand where it fits.
Rather than starting with the technology and looking for places to deploy it, start with the experience you’re trying to create. Then work backward through the processes, systems, people, and decisions needed to make it possible.
It sounds straightforward. In a large organization, it rarely is. But that is precisely why the operating model deserves as much attention as the technology itself.
We’ve spent years improving how customer service operates within the boundaries of the contact center. The next generation of CX will require us to look much further beyond those boundaries. It will depend on how well our systems connect, how effectively our teams collaborate, and whether we’re willing to change processes that have been in place for years.
AI might accelerate that change. It won’t make the organizational decisions for us. And perhaps that’s the question CX leaders should be taking into their next planning cycle.
Are we building a better way to handle customer problems, or are we finally building a business that creates fewer of them?
Guest blog post written by Ruth Rose, SVP of Enterprise Growth & CX Transformation at Transcom.
