By Christian Hasselström
Global Director
Client Success and Solutions – Travel and Hospitality

Industry experts projected that by 2030, the cost of resolving customer service and customer experience (CX) issues using generative AI (GenAI) would exceed the cost of human outsourced customer support agents. This challenges a persistent assumption that adopting AI and GenAI for CX would naturally deliver cost advantage.

Over the past years, organizations have “AI-fied” their CX with efficiency as the primary goal. Chatbots expanded across channels, copilots permeated human workflows, and intelligent automation promised faster responses at lower cost. More automation would mean less effort, fewer resources, and improved economies of scale.

Operational reality, however, has proven more complex. In fact, analysts expect that next year, over 50% of enterprises will reverse earlier plans to reduce their customer service teams. Many are realizing that AI-driven efficiency isn’t enough. While individual touchpoints might become faster, the overall effort required to reach resolution for customers often remains unchanged or challenging. How can CX leaders sustain value once the cost advantage of AI is no longer guaranteed? 

The companies that win in CX-led AI will not necessarily be those with the best models, but those with the lowest cost per resolution and the quickest resolution.

Why measuring cost per interaction misrepresents ROI in AI for CX

Today’s CX operates as a connected journey rather than a series of isolated interactions. A chatbot interaction can become an email thread, escalate to a live human agent, and might require further follow-up when context does not carry forward. Behind the scenes, different tools and teams manage each step, often optimized around separate metrics rather than shared outcomes. When interactions are disconnected, every new touchpoint adds work — and cost — to the journey. Let’s extrapolate from recent industry benchmarks, based on single customer interactions that typically span five to seven touchpoints:

As the numbers show, an omnichannel customer journey can accumulate up to US$26 per interaction. While AI can be present throughout the journey, the overall service economics remain largely unchanged. Many organizations still evaluate AI and GenAI for CX through cost per interaction, even though CX spans channels and teams. AI can reduce the cost of a step, but without coordination across the journey, it does not fundamentally change the cost of resolution.

How orchestration can be the enterprise’s strategic advantage for AI and GenAI

If AI alone no longer guarantees lower cost or differentiates a brand, outcomes depend on how technology, data, and human expertise operate together. Adoption shifts from which AI or GenAI an organization uses to how intelligently it’s embedded within the broader system of customer service delivery.

This is where orchestration becomes a strategic advantage in AI for CX. It connects AI capabilities with operational workflows, allowing systems, teams, and channels to function as a coordinated process rather than a collection of independent tools. In fact, orchestrated omnichannel customer journeys have been shown to improve execution efficiency by 20%, reducing the number of interactions and customer effort required altogether. 

In another example, orchestration in agentic AI for CX enabled a restaurant technology company to achieve a 40% improvement in resolution rates for chatbot inquiries. Rather than simply accelerating responses, orchestration helped resolve issues earlier in the journey.

Let’s adapt these improvements in efficiency (20%) and resolution rates (40%) to our earlier numbers. Rather than discounting interaction costs, orchestration reduces escalation frequency and lowers the effective cost contribution of downstream touchpoints:

While the modeled figures show lower average spend, the improvements come from measuring cost per resolution rather than counting individual interactions.

We’ve achieved similar results in real-world CX operations. In one deployment, TDCX integrated TED, our GenAI-powered agent assist solution, into the workflows of a global travel and mobility company. Agents previously spent as much as 40 minutes assembling information across multiple systems to craft a single response. By bringing knowledge access, case history, and guided responses into one orchestrated environment, response time dropped to just one minute. Productivity increased by 39%, and frontline support agents resolved 33% more cases each day.

How enterprises should measure ROI in AI, GenAI, and agentic AI for CX

Many companies still assess AI performance with efficiency metrics designed for earlier customer service models. They might explain how interactions are handled, but reveal little about whether customers achieve their intended outcome. A “resolved case” can still represent unnecessary effort — when customers repeat the same information multiple times, or only after avoidable handoffs and escalations.

Consider an airline disruption. Resolving an inquiry efficiently through a chatbot or human agent reduces handling effort, but the traveler still needs to contact support. When systems detect disruption early, notify the customer proactively, and offer options before contact is initiated, the customer’s concern is addressed promptly before they need to reach out. 

This shift reflects orchestration at work: coordinating data, AI-driven decisions, and operational processes so that resolution happens upstream:

Why AI adoption must shift toward orchestrated execution 

AI for CX has largely been optimized around interactions — faster responses, higher containment, and shorter handling times. These are important, but when costs are figured into the equation, the discussion shifts from whether AI is becoming more expensive to whether performance is measured where value is created. Organizations realizing meaningful ROI are not necessarily deploying more advanced models, but are harnessing the synergy of AI, data, and human expertise.

Before introducing another AI tool, CX leaders should instead ask if it merely accelerates activity or transforms how customer service is delivered:

• Can systems access and retain context across prior touchpoints?
• Does AI reduce the need for customers to contact support, or simply handle contacts faster?
• Do frontline support agents have visibility into the full history of AI-assisted interactions?
• Are decisions coordinated across channels, or optimized in isolation?

AI, GenAI, and agentic AI can expand what organizations can do, but orchestration turns those capabilities into business value. This sits at the center of a supercycle where technological, operational, and regulatory shifts are redefining CX’s scale, speed, and performance. TDCX’s global delivery model across CX consultancy and CX outsourcing, workforce strategies, and embedded governance is strategically poised to enable organizations to move beyond AI adoption toward thoughtful execution.