Poor AI memory is pushing Asia-Pacific consumers away from brands, as customers grow frustrated with AI-powered service agents that fail to recognise them or remember previous interactions, according to new research by cloud communications company Twilio.
The 2026 Customer Insights Series study found that consumers increasingly encounter “AI amnesia”. This refers to AI agents lacking context about who the customer is or what was communicated in previous interactions.
As a result of this the poor service, some 70 per cent of Asia-Pacific consumers have stopped using an AI agent mid-interaction. This also resulted in 5 per cent of customers ending their relationship with a brand.
The findings reveal that even though AI agents are taking over a significant share of customer operations, the pace of automation adoption is exceeding that of trust.
The Twilio study showed a widening gap between how businesses believe their AI systems are performing and how consumers experience them.
Some 84 per cent of Asia-Pacific brands believe their AI agents do recognise returning customers and their history, but 70 per cent of consumers say they often have to start from scratch when in a new conversation with an AI agent.
The issue continues when the customers are transferred from an AI agent to a human. Here, 65 per cent of Asia-Pacific consumers report having to repeat themselves or fill in information gaps.
According to Twilio, both human and AI agents, are unable to remember a customer when their history is spread across disconnected systems.
“AI is only as helpful as the context you feed it,” said Robert Woolfrey, vice president, Asia Pacific and Japan, at Twilio.
“By connecting customer data, brands can evolve from basic chatbots to smart assistants that remember every individual,” he noted.
He added that repeated explanations should become less common as brands build continuous conversations that allow customers to resume where they left off.
Trust becomes critical with growing AI use
Trust is an important consideration with the growing AI adoption. The Twilio study found that 70 per cent believe AI agents should identify themselves when a conversation begins, but just 22 per cent of Asia-Pacific businesses currently do so.
Consumers also seek greater transparency and control, which includes the ability to request a switch to a human agent (58 per cent), 57 per cent want to know that AI actions require human approval, and 51 per cent want visibility into the customer data that AI can access.
Businesses are starting to respond. The study found that 47 per cent of businesses are prioritising explainability features, 44 per cent are making it easier for customers to reach a human agent, and 39 per cent are giving customers more control over whether their data is used to train AI models.
“AI use will only transform customer experience if it earns trust first,” said scientific futurist Dr Catherine Ball.
“Transparency matters, and businesses that are clear about when AI is being deployed will be better placed to build stronger, and more genuine customer relationships,” she added.
Dr Gemma Calvert, professor of consumer neuroscience at Nanyang Technological University, said that expectations are central to trust.
“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,” she said.
She added that brands need to be clear about what AI can and cannot do, and to disclose when it is being used, and provide an easy route to a human before frustration escalates.
Consumers embracing AI-to-AI
Despite the frustrations with AI memory, consumers are increasingly open to AI-first engagements.
The Twilio study found that 68 per cent of Asia-Pacific consumers agree that the AI bots they interacted with have improved in the past year. Some 65 per cent want AI agents from different companies to communicate directly with one another to resolve issues faster.
Many customers are also comfortable allowing AI to carry out everyday tasks. These include scheduling appointments for 86 per cent, making dinner reservations for 85 per cent, processing returns for 85 per cent, and choosing concert seats for 82 per cent.
Businesses are also preparing for this more AI-led future. Asia-Pacific business leaders expect AI agents to handle 65 per cent of all customer service interactions by 2027, an increase from 52 per cent today.
In addition, there is a shift away from purely reactive support. Some 91 per cent of business leaders are actively building AI workflows to enable proactive customer support before customers reach out, according to the Twilio study.
