Seven in 10 consumers across Asia Pacific have abandoned an AI customer service chat because the system forgot who they were or what they had already discussed, according to Twilio's 2026 Customer Insights Series. The consequences are immediate and commercial-5% of those respondents stopped using the brand entirely after a single failed exchange.
The research, which surveyed 7,652 consumers and 660 business leaders across 18 global markets, exposes a sharp disconnect between corporate confidence and actual customer experience. While 84% of businesses believe their AI agents recognise returning customers and retain conversational history, 70% of consumers said they regularly have to start from the beginning when opening a new query.
Customer history is getting lost between systems
Handoffs from AI to human agents rarely preserve what was already discussed. Twilio found that 65% of consumers have had to repeat information or fill in missing details after a bot transferred them to a person. The root cause is customer histories scattered across disconnected platforms, leaving both AI and human agents without the context needed to continue an earlier conversation.
"AI is only as helpful as the context you feed it. By connecting customer data, brands can evolve from basic chatbots to smart assistants that remember every individual. Having consumers repeat their issues will become a thing of the past as customers have continuous conversations that pick up right where they left off," said Robert Woolfrey, Vice President, Asia Pacific & Japan, at Twilio.
For support teams looking to close this gap, AI for Customer Support training can help teams design agents that pull from unified data sources rather than isolated systems.
Consumers want clearer disclosure and control
Transparency expectations are high and largely unmet. Some 70% of Asia Pacific consumers believe AI agents should identify themselves at the start of every conversation, yet only 22% of businesses currently do so. Access to human oversight remains a priority: 58% want the option to switch to a person on request, 57% want AI actions to require human approval, and 51% want to know what customer information an AI system can access.
Businesses are beginning to respond. Twilio's data shows 47% are prioritising features that make AI decisions easier to explain, 44% are making it simpler for customers to reach a human agent, and 39% are giving customers more control over whether their data is used to train AI models.
Dr Gemma Calvert, Professor of Consumer Neuroscience, warned that poor AI experiences become indistinguishable from brand perception. "Trust depends heavily on expectations. One lesson from the early chatbot era is that when AI is deployed before it can cope with the range and complexity of real customer needs, consumers do not simply blame the technology; that failure becomes part of their experience of the brand itself. Brands therefore need to be clear about what AI can and cannot do, disclose when it is being used, and provide an easy route to a human before frustration escalates," she said.
Businesses expect AI to handle more customer service
Comfort with AI is growing despite context and transparency concerns. Some 68% of Asia Pacific respondents said the AI bots they interacted with had improved over the past year, while 65% want AI agents from different companies to communicate directly when doing so could resolve an issue faster. Many consumers are also willing to assign routine tasks to AI-86% would use it to schedule appointments, 85% for dinner reservations, 85% for processing returns, and 82% for choosing concert seats.
Commercial reliance on automated service is set to expand. Asia Pacific business leaders expect AI agents to handle 65% of customer service interactions by 2027, up from 52% today. Twilio also found that 91% of organisations are already building AI workflows designed to provide proactive support before a customer ever contacts the business.
Why this matters for customer support professionals
The data makes one point clear: customers will not tolerate AI that forgets them. The 5% brand abandonment figure translates directly into revenue loss, and the 65% repeat-information rate signals broken handoff processes that waste agent time and erode trust. Support leaders who unify customer data across systems before scaling automated channels will see fewer escalations and lower churn. Those who skip that step will keep losing customers over conversations their bots should have remembered. For supervisors managing these transitions, AI for Call Center Supervisors provides practical frameworks for overseeing hybrid human-AI teams without sacrificing service quality.
Your membership also unlocks:
