The Hidden Cost of ‘AI Amnesia’: Why Fragmented Data is Eroding Customer Trust in APAC

Imagine a customer in Jakarta spending 20 minutes detailing a billing dispute to a chatbot, only to be transferred to a human agent who requires the entire story to be repeated from the beginning. When scaled across the millions of AI-mediated interactions occurring daily throughout Asia Pacific, this pattern reveals a growing trust deficit beneath the region’s rapid AI customer service expansion.

According to the 2026 Customer Insights Series from customer engagement firm Twilio, consumer frustration is backed by significant data. The report identifies a phenomenon termed “AI amnesia,” where AI agents fail to recognize returning customers or retain context from previous parts of a conversation.

Research shows that 70% of APAC consumers have abandoned an AI-driven service interaction because the system lacked necessary context or failed to recognize them. More critically, 5% of consumers responded to these failures by abandoning the brand entirely.

As AI adoption accelerates across the region—from Singapore’s financial institutions and Indonesia’s e-commerce leaders to the Philippines’ massive BPO sector—this gap between technological ambition and effective execution is becoming an expensive liability.

A widening perception gap between brands and users

There is a stark disconnect between how companies perceive their technology and how customers actually experience it. While 84% of APAC brands believe their AI agents effectively recognize customers and recall history, 70% of consumers report having to start every new interaction from scratch. This discrepancy suggests that many product teams may be relying on flawed internal metrics.

The issue persists even during human hand-offs; 65% of consumers state they must repeat information when transitioning from a bot to a human representative, indicating a systemic data problem rather than a simple chatbot error.

Robert Woolfrey, Twilio’s Vice President for Asia Pacific and Japan, notes that an AI’s utility is strictly limited by the context it receives. He argues that brands can evolve from basic chatbots into true digital assistants only by successfully integrating their customer data.

Twilio identifies the root cause as data fragmentation. Customer histories are often scattered across disconnected CRMs, support ticketing systems, and loyalty applications. Without a unified data layer, neither humans nor AI can construct a complete, coherent view of the customer.

This fragmentation is particularly acute for Southeast Asian businesses operating modern AI agents alongside legacy systems that were never designed for real-time integration. This challenge provides a significant opportunity for the next generation of AI-native CRM solutions.

The demand for transparency and control

Beyond memory, transparency is a major driver of consumer trust. The report finds that 70% of consumers expect AI agents to identify themselves clearly at the start of an interaction, yet only 22% of APAC businesses do so consistently. This suggests many brands view AI disclosure as an optional courtesy rather than a commercial necessity.

Consumer expectations are becoming increasingly sophisticated. 58% of users want the option to request a human agent, 57% want assurance that AI actions require human oversight, and 51% desire visibility into what specific data the AI can access. The central question for consumers has shifted from “does the bot work?” to “can I control what the bot knows about me?”

While some businesses are addressing these concerns by prioritizing explainability (47%) and easier human escalation (44%), a significant gap remains between consumer expectations and corporate delivery.

Gemma Calvert, a professor of consumer neuroscience, warns that when AI is deployed before it can meet fundamental customer needs, consumers do not view it as a mere technical glitch; they view it as a failure of the brand itself. For companies rushing to ship products, a “forgetful” chatbot is not just a UX bug, but a long-term threat to brand reputation.

The transition toward AI-to-AI service models

Despite current friction, APAC consumers are not rejecting AI; in fact, 68% believe the quality of AI service has improved over the last year. Remarkably, 65% of consumers are comfortable with the idea of different companies’ AI agents communicating directly with one another to resolve issues without human intervention. This openness is a vital precursor to the agent-to-agent commerce model currently being developed in the region’s fintech sectors.

Task delegation to AI is already becoming mainstream. Consumers show high levels of confidence in letting AI handle appointments (86%), dinner reservations (85%), processing returns (85%), and even selecting concert seats (82%). These figures demonstrate a level of trust in AI judgment that was unthinkable just a few years ago.

Business leaders are preparing for even deeper integration. In APAC, it is expected that AI agents will handle 65% of all customer service interactions by 2027, up from 52% today. Furthermore, 91% of leaders are building workflows for proactive support, aiming to resolve issues before the customer even reaches out.

Strategic implications for Southeast Asian innovators

For the region’s fast-growing fintech, e-commerce, and customer engagement sectors, the Twilio report provides a clear mandate: the race to deploy AI must be matched by a race to unify data and ensure transparency. Solving the “memory” problem cannot be achieved through isolated software patches on a fragmented infrastructure.

As agent-to-agent communication becomes the norm, the enterprises that prioritize unified customer data and honest AI disclosure will secure a “trust dividend.” In a market where digital reputation can be won or lost instantly, fixing AI’s memory may be just as vital to commercial success as the underlying artificial intelligence itself.

Source link