Consumers in Asia Pacific might be increasingly open to the use of artificial intelligence (AI) in their interactions with brands, but a 2024 PwC survey revealed that they are not offering AI a blank cheque.
More than half of those surveyed indicated they are comfortable allowing AI to manage low-risk interactions, such as product recommendations or basic customer service. However, when the stakes are higher, such as receiving a medical diagnosis or completing a financial transaction, they still prefer having a human behind the interaction.
This boils down to the fact that consumers want to feel heard, understood, and supported, particularly when the situation involves more complexity or nuance. Many would agree that humans do this best.
This presents a dilemma for businesses seeking to harness the power of agentic AI. There is no doubt that AI agents – equipped with reasoning, memory, and orchestration capabilities – have unlocked new levels of productivity and efficiency. They can autonomously manage multi-step workflows, assist with decision-making, and free teams from repetitive tasks. But as organisations race to adopt these technologies, one question remains: how can they tailor AI-first services that drive meaningful outcomes without losing the human connection?
While agentic AI holds immense potential to take customer experience (CX) to the next level, it is clear that organisations must strike the right balance when it comes to when and where to deploy agentic AI. If not done thoughtfully, the use of AI could risk doing more harm than good to consumer satisfaction and brand loyalty.
What exactly does not too much, not too little look like?
See also: Apple, Google warned by Germany that DeepSeek is illegal
When to automate
According to a report by McKinsey, the CX landscape in the region has evolved to be omnichannel, especially amongst the younger generation. Consumers expect brands to deliver seamless, personalised support across all touchpoints, whether via live chat, email, social media, phone, or video.
To ensure they can deliver a high-quality, consistent experience across these different channels, many forward-looking businesses have adopted agentic AI solutions that incorporate smart self-service, live agent support, and contextual handoffs in a single interface.
See also: CoreWeave CEO joins world’s richest after AI firm’s 300% surge
For instance, when a customer reaches out, AI can retrieve prior interactions, anticipate the likely reason for the outreach, and suggest appropriate resolutions. Additional features such as live transcription and conversation summaries also help maintain continuity when issues are escalated or transferred. This ensures that customers get the speedy and accurate resolutions they want, without having to explain their issues multiple times as they move from one channel to another.
Agentic AI also has a critical role to play behind the scenes, particularly in light of rising agent burnout. CMP Research reports that among agents experiencing high levels of stress at work, more than 53% are likely to leave their organisations within six months.
Beyond automating repetitive tasks such as form-filling or basic troubleshooting, agentic AI can provide real-time guidance, offering live prompts, recommended responses, or next-best actions during customer interactions. This helps agents stay focused, reduces cognitive load, and enables them to deliver better outcomes by streamlining their work.
As CX touchpoints expand and become increasingly complex, it is up to businesses to identify gaps in their customer journey and question whether these can be resolved effectively by introducing an autonomous system. If the answer is yes, it is time to automate.
Understand where human empathy still matters
Sometimes, the answer could very well be no.
As the PwC survey highlights, not all customer interactions should be automated. Certain situations require a significant level of emotional intelligence and empathy, elements that only a human agent can provide.
To stay ahead of the latest tech trends, click here for DigitalEdge Section
These include cases involving distress, uncertainty, or high personal stakes, such as billing disputes, service failures, or potential data breaches. In such circumstances, customers are not simply looking for solutions. Instead, they are seeking reassurance, clarity, and a sense of being cared for.
Agentic AI still has a role to play in these instances, but as a supporting tool. Consider the case of a customer who has just missed a flight. AI can quickly aggregate data from multiple systems – past bookings, flight rules, and prior case resolutions – to generate potential next steps. The human agent, informed by this context, can then focus on problem-solving while delivering an experience that is emotionally attuned to the customer’s distress
From Human vs AI to Human-AI Partnership
CX has become the next battleground for organisations in Asia Pacific seeking a competitive edge. Agentic AI can indeed process vast amounts of data to determine the “what” and “when” of customer interactions, but it is the human element that addresses the crucial “why”. Achieving this balance between the tactical and the emotional is what transforms CX from transactional exchanges into meaningful, trust-based relationships.
That said, the road to finding this optimal interplay rests on a very human ingredient: the ability of organisations to cultivate an AI growth mindset in their customer service teams. Rather than viewing AI as a replacement for their skills, human agents should be encouraged to see it as an evolving tool that can augment performance and adapt to changing needs.
Ultimately, companies that treat agentic AI as a mere tool for automation will struggle to earn customer trust, while those who deploy it with purpose (and in the right areas) will set a new experience benchmark in the AI era.
Not too much, not too little – that is the agentic AI sweet spot.
Lukas Carruthers is the head of CX for APJ at Zoom
