The first wave of AI adoption was defined by experimentation. Businesses rushed to deploy chatbots, test generative AI tools and launch automation projects, often without a clear view of the problem they were trying to solve. Service leaders were told they needed “an AI strategy”, often before the business case or outcomes were clear.
The second wave looks very different. Rather than asking where AI can be deployed, organisations are increasingly asking where it can create measurable business value. In customer service, that shift is transforming AI from a standalone technology initiative into a catalyst for broader operational change.
Organisations are now moving beyond experimentation and beginning to treat AI as a foundational business capability rather than a standalone technology layer.
“We’re now in a really interesting stage where businesses are starting to think much more holistically,” says Lucinda McCaffrey, AI sales manager at Zendesk. “Instead of asking, ‘How can I use AI?’, they’re asking, ‘What are the pain points in my organisation, and how can AI remove them?’”
That shift is creating a clearer connection between AI investment and business outcomes across industries.
Moving beyond the AI rush
Early AI projects focused heavily on automation and cost reduction. Chatbots were introduced to deflect customer queries and reduce pressure on contact centres. Success was measured largely through efficiency metrics. But many organisations quickly discovered that deploying AI without a broader operational strategy produced limited results.
“There were quite a lot of assumptions about what AI could and couldn’t do,” says McCaffrey. “People were trying to build business cases around the technology itself rather than focusing on the outcomes they actually needed.”
The question is no longer whether an organisation has an AI strategy – it’s whether its operating model is ready for AI.
Today, organisations are taking a more mature approach. Rather than treating AI as a layer added onto existing processes, many are redesigning workflows, data structures and service models around AI capabilities from the outset.
“We’re seeing businesses think much more seriously about what being AI-ready actually means. That includes governance, data readiness and systems optimisation,” explains McCaffrey.
The pressure driving that change is intensifying as customer expectations grow. According to Zendesk’s latest CX Trends research, customers increasingly expect faster, more accurate resolutions and are becoming less tolerant of poor service experiences.
From cost reduction to transformation
One of the biggest changes McCaffrey sees is how organisations define value from AI initiatives.
“AI projects are moving away from assumed value into very specific and tailored value,” she says. “An AI chatbot can do so much more than just deflect contact.”
An AI chatbot can do so much more than just deflect contact
Rather than pursuing large-scale workforce reductions, many businesses are using AI to remove repetitive, low-value work and allow employees to focus on more strategic activities.
“The staffing strategy for AI is changing. Businesses are saying, ‘Actually, we don’t necessarily want fewer people – we want people working on better optimisation projects inside the business,’” says McCaffrey.
The shift is already visible among organisations moving beyond efficiency-led AI projects. At Fortnum & Mason, the retailer deployed AI not simply to deflect customer contacts but to better understand customer intent and determine when human intervention adds the greatest value.
“I love that the AI really understands the tone and context,” says Rhonda Floyd, head of customer success at Fortnum & Mason. “It knows when to escalate a ticket to a live agent – that way, our agents can really focus on customers with more complex interactions.”
The result was a 90% reduction in live-chat processing times and a 75% fall in average ticket handling times.
Fragmented systems aren’t a blocker for AI-literate companies
One of the long-standing challenges for service transformation has been fragmented technology environments. Service teams often rely on multiple systems spread across departments including finance, logistics, HR and customer support.
Historically, employees were required to act as the integration layer between systems.
“In the past, businesses relied on human agents to become experts in 10, 15 or even 20 different systems,” says McCaffrey. “That obviously isn’t efficient.”
But advances in AI connectivity, APIs and agentic AI models are now making it easier to orchestrate tasks across systems without requiring employees to manage those interactions manually.
The education software provider Arbor offers a good example. Supporting thousands of schools requires teams to navigate large volumes of operational and customer data spread across multiple systems. By centralising support operations and introducing AI-driven workflows, Arbor reduced the need for employees to manually bridge those systems, allowing support teams to scale more effectively.
“Fragmented systems are still a problem, but they’re no longer a blocker,” adds McCaffrey.
In this vein, the cryptocurrency financial services company, Blockchain, is increasingly managing complex regulatory and compliance requirements by implementing strict data access controls within its AI Agent infrastructure. By deploying AI across its messaging channel, the company deflects over 51% of inquiries, effectively automating more than half of its monthly tickets. The AI tools are now enabling quicker resolutions for common issues and provides enhanced interactions when cases are escalated to human agents. This strategy has reduced average agent handling time by 25%, while AI-assisted escalations now streamline 50% of cases requiring human intervention.
Industry transformation is accelerating
While AI adoption is accelerating across the board, some of the most significant progress is occurring in sectors that have historically been difficult to automate. Improvements in reliability, governance and accuracy are making AI increasingly viable in environments where compliance, regulation and complex data requirements once limited its use.
Healthcare is a prime example. The sector has traditionally struggled to automate service interactions because of the volume of patient information, administrative complexity and sensitivity involved.
“We’re unlocking industries that weren’t necessarily there before,” says McCaffrey.
At the same time, smaller businesses may have advantages of their own. Many newer organisations were built in cloud-native environments with fewer legacy systems and more flexible operating models.
“This is almost the time for the little guy. Some newer businesses are already built for this agentic era,” says McCaffrey.
One of the most important changes underway is the shift from reactive customer support towards resolution-driven service models. The goal is no longer simply responding to queries faster, but resolving issues more effectively and with less effort for customers and employees alike.
“People need to go back to basics,” she says. “Map the customer journey from start to finish and think about what the customer actually expects at every point.”
That can include personalised recommendations, proactive support and intelligent workflow orchestration across different stages of the customer lifecycle. As the impact and ROI of AI evolves, businesses are also moving beyond simple deflection metrics.
“Deflection was really just measuring whether AI worked at all. Now organisations are looking much more at customer satisfaction, employee satisfaction and total cost of ownership,” says McCaffrey.
For organisations still early in their AI journey, McCaffrey’s advice is simple: focus less on the hype and more on understanding where AI can create meaningful value.
“AI isn’t just a business problem, it’s an individual problem too,” she says. “People need to understand what this technology means for them, their customers and the businesses they work in.”
Find out more about AI-powered service transformation with Zendesk
The first wave of AI adoption was defined by experimentation. Businesses rushed to deploy chatbots, test generative AI tools and launch automation projects, often without a clear view of the problem they were trying to solve. Service leaders were told they needed “an AI strategy”, often before the business case or outcomes were clear.
The second wave looks very different. Rather than asking where AI can be deployed, organisations are increasingly asking where it can create measurable business value. In customer service, that shift is transforming AI from a standalone technology initiative into a catalyst for broader operational change.
Organisations are now moving beyond experimentation and beginning to treat AI as a foundational business capability rather than a standalone technology layer.
