Asia’s Next AI Revolution: Two IDC Forecasts, a $32 Billion Opportunity and the Race to Transform Customer Experience

IDC predicted that 60% of Asia-Pacific’s largest companies would redirect 35% of their customer service budgets towards emerging products and skills by 2027. It also forecast US$32 billion in consumer spending through AI agents by 2028. As these deadlines approach, one question matters more than either prediction: will AI actually make doing business easier for customers, or will it simply make businesses more expensive to run?
A customer contacts a bank because a transaction has failed. The AI understands the problem. It responds instantly. It sounds empathetic. It may even explain why the transaction failed.
But it cannot access the relevant transaction record, cannot initiate the required action and cannot tell the customer when the money will be returned. The customer eventually speaks to a human agent, repeats the entire problem and waits again. The company reports another successful AI interaction. The customer remembers another frustrating experience. This is the uncomfortable gap between the AI transformation businesses are announcing and the customer experience they are actually delivering.
And it is exactly why two IDC predictions about Asia-Pacific deserve a closer look as we approach 2027. The first concerns how much businesses will invest in transforming customer service. The second concerns how much consumers could spend through autonomous AI agents. Together, they point towards a massive commercial opportunity. But they also raise a question that the industry cannot afford to avoid:
If AI is becoming the next interface between businesses and customers, who is ensuring that the interface actually delivers value?
The first prediction: 35% of customer service budgets could be redirected. But what will that money achieve?
In January 2024, IDC’s Asia/Pacific FutureScape predictions indicated that 60% of the region’s top 2,000 companies would reallocate 35% of their customer service budgets towards emerging products and skills by 2027.
This was not a prediction that 35% of budgets would go exclusively to generative AI. It was a broader forecast about investment in emerging technologies and capabilities. But the message was significant: customer service was heading towards a major technology-led transformation.
Now look at the spending signals. Gartner’s August 2026 research found that customer service and support leaders had increased AI spending by 38%, while overall service and support budgets had grown by just 2%. That is a striking difference. AI is receiving a substantially larger share of management attention and investment.
But spending is only half the story. Gartner’s July 2026 research found that just 24% of surveyed customer service leaders demonstrated positive financial returns across their AI use cases. The median share of their 2025 budgets invested in AI was 12%.
Put these findings together and a very different picture emerges. AI spending is accelerating. Yet positive financial returns are not widespread among the leaders surveyed. This is the first mistake enterprises must avoid: confusing technology investment with business transformation.
A company can invest in generative AI, introduce virtual assistants, automate customer conversations and announce an ambitious AI roadmap. None of those activities automatically proves that customers are receiving better service or that the business is operating more efficiently.
The real question is not how much of the customer service budget goes to AI. It is how much customer effort, operational cost and unresolved demand that investment removes.
For businesses evaluating Conversational AI & Customer Experience Summit Asia priorities, this is where the conversation needs to move: from AI adoption percentages to measurable customer and commercial outcomes.
The second prediction: US$32 billion. What happens when AI becomes the shopper?
The next prediction is even more interesting because it moves AI beyond customer service and into commerce. In February 2025, IDC forecast that consumers across Asia/Pacific, excluding Japan, would spend US$32 billion through autonomous AI agents by 2028, as agents independently shop for goods and services.
Consider what that could mean for a retailer. Today, a customer might search for a product, visit several websites, compare specifications, read reviews, check prices and finally complete a purchase. Tomorrow, the customer might simply tell an AI agent:
“I need a laptop for video editing. My budget is ₹80,000. Compare the best options, check availability, find the best deal and ask me before purchasing.”
The agent could handle much of the discovery and comparison process. With the right integrations and permissions, it could also help execute the transaction. This changes something fundamental.
For years, brands competed to attract customers to their websites, apps and sales channels. In an agent-led shopping environment, brands may increasingly compete to be understood, evaluated and recommended by an AI intermediary.
The customer journey could begin before a customer ever visits a company’s website. That has major implications for conversational commerce, product information, customer data, payments, personalisation and brand visibility. And the adoption signals are already emerging.
Deloitte’s April 2026 report, The Future of Commerce: Agentic Shopping in Asia Pacific, found that:
- 29% of surveyed Asia-Pacific consumer businesses were already adopting agentic AI.
- 76% expected to adopt it within two years.
- Almost three-quarters of consumers surveyed were using AI to discover, compare and learn about products.
- Only around 30% of consumer businesses reported that at least 40% of their AI initiatives reached production.
The final figure is particularly important. Businesses are preparing for agentic commerce, but getting AI initiatives into production remains a challenge. The opportunity is not simply to build an AI shopping assistant. It is to build a business that an AI shopping assistant can actually transact with. That requires accurate product data, real-time availability, reliable integrations, transparent pricing, secure payment processes and clear transaction permissions.
If an agent recommends a product that is unavailable, quotes an outdated price or cannot complete the purchase, the experience breaks. The US$32 billion prediction is still a future forecast, not a confirmed result. But it gives businesses a compelling reason to prepare for a customer journey in which AI could become an increasingly important decision-maker.
The Uncomfortable Customer Reality: Your Chatbot May be Losing to Someone else's AI
There is another development that deserves much more attention from customer experience leaders. Gartner’s July 2026 survey of 3,566 B2B and B2C customers found that customers were approximately three times more likely to use third-party generative AI tools than company-provided chatbots when resolving customer service issues.
Think about the implication. Businesses have spent years building their own digital assistants. Yet customers may prefer asking an external AI tool for help with a company’s products and services.
Why?
Because customers are not loyal to a chatbot. They are loyal to convenience.
They do not care whether a business uses a sophisticated language model, a proprietary assistant or an advanced conversational interface. They care about whether their problem gets solved without unnecessary effort.
A customer checking a delayed delivery wants an accurate update, not a link to a tracking page they have already visited. A customer disputing a payment wants an investigation, not another explanation of the refund policy. A customer changing a subscription wants the change completed, not instructions explaining how to navigate the account settings. This is where traditional chatbot strategies reach their limits.
Understanding customer intent is not the same as fulfilling customer intent. For AI to deliver meaningful customer experience improvements, it must connect to the systems and workflows that allow businesses to take action.
That includes CRM platforms, order management systems, payment infrastructure, ticketing tools, customer databases and relevant business applications. Without those connections, even an intelligent AI assistant can become another layer between the customer and the solution.
For us at Conversational AI & Customer Experience Summit Asia, this is one of the most important distinctions in the evolution of conversational AI: the objective should be to move beyond conversations and enable meaningful customer outcomes.
Agentic AI in Customer Service: The biggest breakthrough may be knowing when not to automate
There is a tendency in the AI industry to celebrate how many interactions an AI agent can handle independently. I think businesses should ask a different question.
Does the AI know when it should not act independently?
Gartner’s August 2026 customer research found that 50% of customers said their interactions were easier when companies used generative AI. Yet 87% said access to a human agent was essential when companies used generative AI in customer service.
That is not resistance to AI. It is a clear statement about what customers expect from a good experience. They want the convenience of automation without losing control. This is why the next generation of Agentic AI in Customer Service must be built around controlled autonomy.
An AI agent should be able to understand a request, retrieve relevant information, execute approved actions and confirm the outcome. But it must also recognise uncertainty, respect permissions and transfer complex cases to a human agent without losing the conversation context.
For example, an AI assistant may be perfectly capable of checking an order and processing an eligible cancellation. But if a customer disputes a payment or raises a sensitive complaint, the system should know when a human decision is necessary.
The goal is not to eliminate human involvement. It is to make human involvement more valuable. This is a more sustainable way to approach AI-driven customer experience because it balances speed, efficiency, accountability and trust.
The metric that could expose the entire AI transformation
If I had to challenge one common measurement in enterprise AI, it would be chatbot containment rate. A high containment rate is often presented as evidence that automation is working. But it only tells part of the story.
What if customers leave the chatbot because it cannot help them?
What if they call the contact centre immediately afterwards?
What if the AI provides an answer, but the customer still has to complete the task manually?
In those situations, the interaction may have been contained, but the problem was not resolved. This is why AI Customer Experience ROI needs a different measurement framework. Instead of asking only how many conversations AI handles, businesses should measure:
- Resolution rate: What percentage of customer issues are genuinely resolved through AI-assisted journeys?
- Cost per resolved interaction: How much does it cost to solve a customer problem, including AI infrastructure, integrations, human support and ongoing maintenance?
- Repeat contact rate: How often do customers return because the original issue was not properly resolved?
- Customer effort: Are customers completing tasks with fewer steps, less repetition and less frustration?
- Conversion and revenue contribution: Does conversational AI help customers make decisions and complete purchases?
- Escalation quality: When AI cannot resolve an issue, does it transfer the conversation with the necessary context?
These metrics connect technology performance to commercial performance. A business should establish a baseline before deploying AI, measure the same outcomes after implementation and calculate the full cost of delivering those outcomes. That is how AI Customer Experience ROI becomes a business case rather than a technology presentation.
Five decisions Asia's enterprise leaders should make before 2027
The next phase of the AI transformation will not be won by the companies that deploy the most assistants. It will be won by organisations that make better decisions about where and how to use them. Here are five priorities I believe deserve immediate attention.
- Stop automating broken processes. Map the customer journey first. Identify where customers experience delays, repetition and unnecessary handoffs. Then decide where AI can remove friction.
- Make customer data usable. Connect relevant systems and information so AI can respond with context. Poor data quality and disconnected platforms will undermine even the most advanced models.
- Move from answers to actions. Prioritise workflows where AI can deliver a measurable result, such as checking order status, updating customer information, scheduling appointments or handling eligible requests.
- Make trust part of the architecture. Establish access permissions, audit trails, escalation rules, security controls and human oversight before expanding autonomy.
- Demand evidence of business value. Set clear targets for resolution, customer effort, cost efficiency, retention and conversion. Review actual performance, not just deployment statistics.
These are also the questions that should shape the Conversational AI & Customer Experience Summit Asia discussion. The industry needs fewer conversations about what AI could theoretically do and more evidence of what it can reliably deliver in real business environments.
Conclusion: The AI winners will be decided by outcomes, not announcements
So, did IDC get its predictions right?
The honest answer is that we cannot yet declare either prediction achieved. The 2027 budget forecast requires a like-for-like assessment of spending across the specified top 2,000 Asia/Pacific firms. Rising AI expenditure is an important signal, but it does not establish that the precise forecast has been fulfilled.
The US$32 billion agentic commerce prediction is for 2028. Current adoption data suggests momentum, but it cannot confirm future consumer spending. What we can say is that the market is moving in the predicted direction. AI investment is increasing. Consumers are using AI in product discovery. Enterprises are exploring agentic workflows. Customer expectations are changing.
But the execution gap remains. At Conversational AI & Customer Experience Summit Asia, we believe the next stage of conversational AI will be defined by a simple but demanding principle: technology must create value beyond the conversation. The businesses that succeed will connect AI to their customer journeys, their operational systems and their commercial objectives. They will know what to automate, what to leave to people and how to prove that the experience has improved.
Because the real competitive advantage is not having an AI agent that can talk to a customer. It is having an AI-powered business that can understand what the customer needs, take the right action and deliver the promised outcome. IDC gave Asia a timeline. The market has supplied the investment. Now enterprises have to deliver the experience.
FAQ’s
IDC predicted that 60% of the top 2,000 Asia/Pacific firms would reallocate 35% of their customer service budgets towards emerging products and skills by 2027. The prediction covered emerging products and skills broadly, rather than generative AI spending alone.
IDC forecast that consumers across Asia-Pacific, excluding Japan, would spend US$32 billion through autonomous AI agents by 2028. The prediction concerns AI agents participating in shopping for goods and services.
Deloitte's April 2026 research found that 29% of surveyed Asia-Pacific consumer businesses were already adopting agentic AI, with adoption expected to reach 76% within two years.
AI Customer Experience ROI can be difficult to demonstrate when businesses deploy AI without connecting it to customer data, operational workflows and measurable business outcomes. High chatbot usage does not necessarily mean more resolved customer issues or lower operating costs.
Agentic AI in Customer Service refers to AI systems that can understand customer intent, retrieve relevant information, make decisions within defined permissions and execute approved actions. Unlike traditional chatbots, agentic systems can help complete customer tasks rather than only provide information.
Businesses should measure first-contact resolution, cost per resolved interaction, repeat contact rate, customer effort, conversion, revenue contribution, customer satisfaction and escalation quality. These metrics reveal whether conversational AI is improving the customer experience and business performance.
Agentic AI can automate repetitive tasks and support customer service operations, but human agents remain important for complex, sensitive and high-stakes situations. Effective customer service combines automation with human judgement and appropriate escalation.
The upcoming Conversational AI & Customer Experience Summit (CACES) 2027 will be hosted across three Asian destinations: Kuala Lumpur, Malaysia, on 15 April 2027; Jakarta, Indonesia, on 15 July 2027; and Singapore on 20 August 2027. Each edition will bring together industry leaders, technology providers, and customer experience professionals to explore the latest developments in Conversational AI, Agentic AI, intelligent automation, and AI-driven customer experience. Event dates and venues are indicative and subject to confirmation.

