
Image generated by Deeptech Times using ChatGPT
AI search is changing how consumers discover brands, products and experiences. Consumers can ask increasingly specific questions and receive a synthesised response drawing on information from across the web.
For marketers, that changes the mechanics of discoverability. The question is increasingly how a brand’s information can be understood, retrieved and surfaced when a consumer asks an answer engine a question.
Raffles Hotel Singapore offers a useful case study. During its 2026 mooncake campaign, the hotel created new website pages and features to strengthen discoverability, including in AI-powered search environments.
Daniel ZX Lim, formerly assistant director of e-commerce and digital strategy, marketing communications at the hotel, described the work as part of a broader digital strategy grounded in consumer behaviour, purchasing preferences, search engine optimisation and answer engine optimisation (AEO). Deeptech Times spoke with him at Tech Week Singapore 2026.
The experiment also exposes a growing problem. A brand can improve its visibility for AI-driven discovery and observe a commercial correlation without a technology stack capable of isolating the contribution of AI.
From search visibility to AI discoverability
AEO is a term for optimising content so answer engines can retrieve and use it when responding to user questions. CXL describes it as making useful information easy for an answer engine to find, understand and use. The discipline overlaps substantially with traditional SEO, particularly around clear information, search intent, authoritative content and technical accessibility.
Raffles Singapore’s approach reflected that overlap. Lim said the new pages and website features were developed from an understanding of the hotel’s target audience, purchasing preferences, consumer behaviour and the strategies that could improve discoverability. The hotel built useful digital information for customers while considering how that information could be discovered in an evolving search environment.
During the campaign, Raffles published detailed information on ordering, storage, delivery and pairings. Its mooncake FAQ guides readers on matching flavours with tea, champagne, wine and sparkling tea. Raffles’s sommeliers curated the pairings according to the characteristics of each mooncake.
“For instance, our sommeliers suggested champagne pairing with our mooncakes as a “luxury enjoyment”,” Lim said.
The content serves several purposes: helping customers make a purchase decision, showcasing the brand’s expertise and creating information that search systems can potentially retrieve when users ask related questions. It also links to products that extend the occasion. Tea is a natural accompaniment to mooncakes, while champagne, wine and sparkling tea add further products to the consideration journey.
The principle is familiar to marketers: understand what customers need to know, answer those questions clearly, and connect the information to a relevant commercial proposition. AI search adds another route through which that information may be discovered.
The opportunity cost of doing nothing
Lim said there may be no definitive evidence showing that one particular page or feature is the correct investment. The alternative is an opportunity cost: the potential missed opportunity of failing to act or lacking the knowledge, skills or budget required to act.
That is a familiar problem in digital marketing. A company rarely has perfect evidence before publishing a product guide, FAQ, buying guide, or other customer-facing content. The decision rests on the strength of the customer need, the commercial opportunity, the available resources and the organisation’s ability to execute.
The gap between AI visibility and AI performance
Raffles Singapore had measurable digital activity associated with the campaign, but according to Lim, those metrics were not recorded in association with AI platform usage.
“No metrics today can show this correlation,” he said. The hotel had no technology that monitored AI marketing performance and connected an AI interaction to a subsequent commercial outcome. Questions about AI-specific traffic, conversions, branded searches or enquiries therefore could not be answered as conventional channel performance questions are.
This is a wider measurement challenge as search becomes more conversational. A consumer could encounter a hotel or product in an AI-generated answer, remember the brand, search for it later, visit the website directly or buy offline. The eventual transaction may contain little evidence of the original discovery. The absence of a measurable AI referral leaves open whether AI had influence, because the current measurement framework cannot reliably quantify it.
Lim said beverage sales doubled year-on-year during the campaign. He attributed the outcome to the wider digital marketing strategy as a whole. Online sales play a major role in Raffles Singapore’s overall mooncake sales each year, although Lim could not provide detail on traffic, conversion rates, average order value or repeat purchases.
The relationship between the campaign and beverage sales is best described as an observed correlation. Lim said other marketing tactics were held at status quo, creating a clear commercial signal during the period. That signal is meaningful for the business, but it does not establish that AI search caused the increase.
At Raffles Singapore, AI advertising sat within the wider digital marketing budget. “Advertising on AI is a very new feature,” Lim said. “So new that not many companies are even able to register or onboard themselves yet.”
He added that the budget covered both SEO and AEO, reflecting the overlap between conventional search optimisation and emerging answer engine practices.
Content still needs a customer purpose
The Singapore campaign ran within a seasonal window, namely, the widely mid-autumn festival. The Raffles Celebrations FAQ gives detailed guidance on storage, delivery, flavour selection and beverage pairings.
The content had an immediate customer purpose independent of AI search. That may be one of the more important lessons for marketers experimenting with AEO.
Content created solely for an emerging algorithm becomes difficult to justify when platforms change. Content that answers a genuine customer question has a more durable role. A customer choosing a flavour may also want to know which drink complements it, and the answer supports discovery, consideration, cross-selling and the customer experience at once.
Lim estimated that AI supported at least half of his e-commerce and digital strategy work, including refining product copy, improving imagery and summarising research. AI-generated copy was reviewed meticulously to keep it within Raffles’ voice and tone before publication.
“As much as AI has made certain tasks easier, the responsibility and impact of the team also intensify in correlation,” Lim noted. The easier it becomes for systems to retrieve and synthesise information, the more important it is that the underlying information is accurate, useful and consistent with the brand.
Raffles’ campaign shows how a brand can create customer-focused content with discoverability in mind, including AI discoverability, while making no claim that every page was created for AI or that AI directly caused a commercial result. The hotel could observe a correlation between the wider campaign and increased beverage sales. It could not isolate AI as a measurable cause within that result.
The question for marketers is increasingly how to make useful information available wherever customers ask for answers, with enough measurement discipline to understand what happens afterwards.











