For years, customer service leaders have talked about moving service from a cost centre to a source of value. But while the ambition was there, the technology, operating model and executive mandate often were not.
That’s now beginning to change as AI forces a rethink of what service can do for the enterprise.
“For many organisations, especially in B2C, this has been an ambition for decades. But until AI became operationally relevant, it wasn’t something they could truly deliver,” says John Kelleher, vice president of enterprise sales at Zendesk.
Kelleher notes that service teams have long been constrained by administrative work and fragmented systems.
“Now with automation and AI really starting to take hold, and freeing up the time of those agents, we can really start to move the dial from it being cost-centric to being more about business change, upsell, cross-sell and more complex customer requirements,” he says.
Beyond automation, AI is exposing how fragmented many service operations have become across customer experience, employee service, contact centres, knowledge bases, workflows and data.
That shift is putting service on the board agenda. CIOs are looking at how to scale AI safely and effectively. CFOs are looking for measurable productivity gains and lower cost to serve. COOs are looking for consistency, resilience and operational visibility. CEOs and boards are asking where AI can deliver value quickly enough to justify the investment.
Increasingly, service is one of the clearest answers.
The hidden cost of fragmented service
However, while the opportunity is significant, so is the barrier. In many enterprises, service is still fragmented by design.
Customer service may sit on one platform and employee service on another. Knowledge is often spread across teams and regions, while AI pilots run in isolated pockets rather than contributing to a broader operating model.
To customers and employees, this fragmentation shows up as friction: repeated information, delayed responses, inconsistent answers and journeys that fail to resolve the underlying issue. To the business, it shows up as a cost.
“The issue with data silos is that it creates process silos – or conversely, it’s the process silo that created the data silo. That’s what creates the fragmentation in the end-to-end process,” explains Kelleher.
This is why service transformation shouldn’t be treated as a technology deployment alone. It requires decisions about operating model, process design, knowledge architecture, governance and change management.
“You can’t look at CX programmes or EX programmes purely as a technology project today,” says Kelleher. “You really need to think about your future-state data and knowledge architecture, your integration strategy, and AI-first process redesign.”
From functional service to enterprise capability
The more strategic opportunity for organisations is to unify service across the enterprise. That means creating a common service layer: shared architecture, consistent knowledge, connected workflows and a single operational view of requests, resolutions and performance.
For large organisations, this can be a significant change. Many have separate service strategies for CX and EX, with different vendors, teams and processes. But Kelleher says Zendesk is seeing more enterprises explore standardisation across both.
“In the last six months, we’ve been having that conversation more. Rather than standardisation within CX, customers are saying, ‘actually, we want to standardise on service across the board’,” he says.
The business case can be immediate. Fewer platforms can mean lower total cost of ownership, reduced vendor complexity and simpler support models. But the bigger prize, arguably, is agility.
When service processes are easier to change, the organisation can adapt faster. When knowledge is connected, AI becomes more useful. When service data is unified, leaders gain visibility into recurring issues, demand patterns and operational bottlenecks.
The cosmetics leader Lush demonstrated the commercial value of this approach by unifying customer care across 21 international markets onto a single Zendesk infrastructure, eliminating fragmented tracking and delivering a 369% ROI in under a year, alongside a 17% increase in agent productivity. This is where service begins to move beyond cost control and become a business enabler.
The ROI is bigger than efficiency
For the C-suite considering the transition, the first question is often financial: where is the return?
The good news is there are clear efficiency gains. AI can reduce cost per interaction, automate high-volume requests and help agents resolve issues faster. A unified platform can reduce duplication and vendor spend. Better workflows can lower handling time and remove manual effort.
But the ROI case doesn’t stop there. Kelleher says many business cases still start with efficiency because that is what CFOs need to approve investment – but once organisations look deeper, the value expands.
“Every customer knows that as we become more efficient from a service perspective, there’s so much more we can do in terms of driving enhanced service levels with our customers, driving upsell with our customers. There’s a growth lever, there’s an NPS lever, there’s a talent redeployment lever,” he says.
That means AI transforms the work service teams are able to do. Routine tasks can be automated, freeing agents to focus on higher-value activities. At Liberty London, Zendesk AI reduced first reply times by 73% and resolution times by 11%, enabling agents to move from ticket sorting into digital personal shopping and styling services.
Crucially, applications such as at Liberty London demonstrate how service agents can become specialists rather than ticket handlers, bringing deeper product knowledge, stronger customer relationships and greater commercial impact.
Why resolution matters
Traditional service metrics often measure activity: response time, handle time, ticket volume, deflection rate. While useful, they don’t always tell leaders whether the customer or employee achieved what they needed.
This shift is also being driven by rising expectations. According to Zendesk’s latest CX Trends research, fast, accurate resolution is a key driver of customer loyalty and purchasing decisions.
A resolution-led model asks a different question: was the issue solved fully, accurately and with the least possible effort?
You can’t just get the immediate impact from AI without thinking about your processes
Embedded into workflows, AI can help classify intent, surface knowledge, recommend actions, automate repeatable steps and support agents through more complex interactions. More advanced agentic AI can take action across systems, moving an issue even closer to resolution rather than simply providing an answer.
At Chiltern Railways, Zendesk’s automated QA and Copilot capabilities helped streamline support for almost 23 million annual passengers. The result was a 40-fold improvement in quality assurance efficiency, a 66% reduction in full resolution times and a 21% increase in complaint-related CSAT.
But driving such results only works when the right foundations are in place. AI needs reliable knowledge. It also needs connected workflows, oversight and be able to operate within a service strategy that defines when automation is appropriate and when human judgement is essential.
“You can’t just get the immediate impact from AI without thinking about your processes,” says Kelleher. “AI-first process redesign means evolving your knowledge, your data architecture and your customers’ readiness for AI transformation.”
Service as the proving ground for AI
One reason service is becoming strategically important is that it offers a practical path to enterprise AI adoption.
Compared with more complex domains such as supply chain or finance transformation, service often has high volumes, repeatable workflows, visible pain points and measurable outcomes. That makes it a natural place to prove the AI operating model before extending it elsewhere.
Kelleher says many organisations are now looking at service in exactly this way.
“As organisations are trying to apply AI at an operational scale, rather than just experimental pockets, they are seeing service as one of the priority domains. They feel they can get the biggest return quicker, almost prove out the operating model and then scale into other functions,” he says.
That has implications for leadership. Service transformation should not sit only with the head of CX or the contact centre. It requires sponsorship from technology, finance, operations and the board.
That’s because the benefits span the different aspects of the enterprise. Cost reduction matters to finance. Architecture and governance matter to IT. Productivity and process resilience matter to operations. And customer loyalty and growth matter to the CEO.
This is where the idea of a resolution platform becomes important as the focus shifts from managing tickets to delivering outcomes.
For business leaders, the point isn’t to add another tool to an already crowded technology estate. It’s to create a strategic layer that connects service demand, knowledge, workflows, AI and human expertise around the common goal of resolution.
A resolution platform gives the enterprise a way to manage service consistently while still allowing different teams to serve different audiences. It supports automation, as well as governance. It enables personalisation, but also visibility. It helps consolidate vendors, but more importantly, it creates a clearer operating model for service.
A new C-suite mindset
The organisations that succeed will be those that stop thinking about service as a downstream function and start treating it as a source of enterprise intelligence.
Every service interaction contains signals: what customers are struggling with, where employees are blocked, which processes are failing, which products are creating avoidable demand, where knowledge is missing, where automation can help and where human expertise is still essential.
Fragmented service hides those signals, whereas unified service makes them visible. For business decision makers, that visibility is becoming critical. In a volatile environment, leaders need to understand where friction is building and how quickly the organisation can respond. Service is often where that friction surfaces first, making it a source of intelligence as much as an operational function.
The transformation frontier, then, isn’t simply about making service faster or cheaper. It is about making the enterprise more responsive, more intelligent and more capable of delivering value at scale.
Find out more about AI-powered service transformation with Zendesk
For years, customer service leaders have talked about moving service from a cost centre to a source of value. But while the ambition was there, the technology, operating model and executive mandate often were not.
That’s now beginning to change as AI forces a rethink of what service can do for the enterprise.
“For many organisations, especially in B2C, this has been an ambition for decades. But until AI became operationally relevant, it wasn’t something they could truly deliver,” says John Kelleher, vice president of enterprise sales at Zendesk.
Kelleher notes that service teams have long been constrained by administrative work and fragmented systems.
