Singapore – Seventy-two per cent (72%) of Singaporeans use AI assistants for shopping, yet only 29% would be comfortable letting one complete a purchase on their behalf, according to a report from Adyen.
It found that consumers and merchants are both adopting AI, but neither group fully trusts it at the point of payment. Among consumers, usage is already routine for many. Some 42% turn to AI assistants for their shopping needs at least once a week.
Willingness to hand over the transaction itself is far lower, however. Some 58% of Singaporeans are uncomfortable with the idea of AI completing a purchase for them.
The hesitation persists among existing users. Nearly half (49%) of those who already use AI for shopping say they would pause at the payment stage.
Merchants show a similar split between performance and confidence. According to the research, AI-referred visitors to retail sites generate 53% more revenue per visit than non-AI traffic, and convert at a 54% higher rate.
Results from AI-driven personalisation are less consistent. Among merchants that deployed AI for personalisation and recommendations, 39% reported a jump in turnover, while 24% reported a drop.
Trust also limits how far retailers are willing to go. Some 36% of merchants cite the risk of losing direct customer relationships or brand control as a reason to delay broader AI adoption.
The Index identifies checkout as a particular pressure point. When payment errors occur, 24% of customers say they may still complete the purchase but would return only if no alternatives exist.
A further 20% would abandon the purchase entirely and avoid the retailer in future, while 15% would switch to a competitor altogether.
“Singapore is one of the most digitally connected markets in the world. Consumers here have embraced AI for discovery and inspiration, and businesses are adopting AI as a reflection of our times,” said Ben Wong, General Manager, Southeast Asia and Hong Kong, Adyen.
“But the data tells a more nuanced story. Adoption is not the same as trust. Hesitation takes over the moment money changes hands — and that gap is exactly where the next chapter of retail will be decided,” Wong added.
Looking at who is driving that adoption, the Index shows usage skewing towards younger, higher-income and male consumers. Gen Z (85%) and Millennials (80%) lead, ahead of Gen X (68%) and Baby Boomers (52%).
Men are more likely than women to use AI shopping assistants, at 76% against 68%. Adoption peaks at 82% among households earning S$8,000 or more annually.
Parents with children under 18 emerge as the heaviest users. They spend more than double per social media purchase compared with those without children (S140againstS67), and use AI more often (82% against 68%).
This group is also more open to AI-completed purchases, at 62% against 54%. The report points to greater time constraints as a possible factor behind their reliance on AI for fast inspiration, personalised recommendations and efficient decision-making.
Even so, rising reliance on AI has not dulled consumer caution. Nearly half of Singaporeans (49%) use biometrics to authenticate transactions, while 29% avoid saving payment details on their devices.
Another 27% feel more confident when retailers use two-factor authentication. Among high-value social media spenders, who spend S$200 or more per purchase, additional security checks nearly double comfort levels, from 11% to 21%.
On the merchant side, awareness of what comes next is high. Nine in ten (89%) say they are familiar with agentic commerce, including 47% who are very familiar and can confidently explain it.
Looking two years ahead, merchants planning to invest in agentic commerce most often cite driving incremental sales through AI-led product discovery (61%). Protecting and retaining direct customer relationships, and managing payment risk and machine-speed transactions, follow closely at 60% each.
“Most payment systems were built for humans, not machines. Agentic commerce brings new rules, new standards and new complexity. The merchants who win will be those who weave trust into the infrastructure itself by making payments invisible, resilient and reliable, at machine speed,” concluded Wong.
