The Gen AI search surge: What marketers need to know

Digital remains the driver of advertising growth for the Asia Pacific (APAC) region. According to the latest report by MAGNA, overall digital advertising revenues are projected to experience a 11.1% year-on-year increase in 2024. Representing US$103 billion in 2024, search commands a 47% share, the largest proportion of total advertising budgets across APAC, as well as in the region’s biggest advertising markets: China, Japan, Australia, and South Korea.

Without a doubt, search continues to be an important marketing tool, and marketers need to keep pace with its evolution.

Search engines are still a primary source for consumers to find relevant information, and in APAC, although Google is the key platform in most markets, local search engines like Baidu in China and Naver in South Korea dominate their respective local landscape. With AI tech now becoming mainstream, all three giants have been quick to build and integrate generative AI functions into their search services. In the last month, Google has rolled out a generative AI search results feature, AI Overview, to more markets globally, including Japan, India, and Indonesia. This offers a more personalised and conversational search experience, and it can respond to complex queries with more-relevant answers.

In India, Google has tailored AI Overview to support both English and Hindi, reinforcing efforts to make AI-powered search accessible to a wider audience. And in a bid that recognises the popularity of voice search, AI Overview also supports text-to-speech, providing AI-generated summaries and responses to voice queries. Going forward, ads will soon appear in a “sponsored section” within AI Overview based on the query’s relevance and information, as testing search and shopping ads in its AI-generated answers are underway.

At the forefront of Baidu’s search revolution is Ernie (known as 文心一言, or Wenxin Yiyan, in Chinese), an AI language model that leverages knowledge graphs to enhance natural language understanding and generation. Its structured database of basic information includes scientific, demographic, geographic, and economic data—as well as, crucially, the semantic links between them. This helps it present facts alongside search results, increasing accuracy and relevance. Baidu is currently on version 3 of Ernie, which debuted in 2019.

In May this year, while Google announced its launch of AI Overview, Baidu disclosed that around 11% of traditional core search results were already being generated using AI technology. It has also been reported that Baidu has been using generative AI to match user queries to advertisements and sponsored search results, helping advertisers to get better at targeting, while also serving users with more-relevant advertising.

Upholding its leadership of the Korean media market, Naver introduced its beta of Cue: a generative AI search engine, in September 2023. Leveraging the internet giant’s extensive language model, HyperCLOVA X, Cue: offers a conversational approach to search queries and is being positioned as an add-on to the current Naver search ecosystem, to avoid detracting from existing results.

What distinguishes Cue: is its integration with various online services within the company’s ecosystem, like Naver Shopping and Naver Place, offering a seamless transition from information retrieval to application and use, and significantly enhancing the search engine’s performance. Integrating search, shopping, reservation, and payment functionalities will further enhance the user journey, eradicating disjointed interactions across various platforms.

That said, dedicated search engines like Baidu and Naver are no longer the only players in search. Depending on what they are looking for, more and more consumers today are turning directly to in-app search or social-media platforms for more-relevant results. It’s therefore unsurprising that the ecommerce giant Alibaba and its social media counterpart Tencent are also leveraging generative AI to level up search in their apps and ecosystems.

Alibaba unveiled Tongyi Qianwen (通义千问), its flagship large language model (LLM), in October 2023. Integrated with its retail business, Taobao, this LLM has been used to create Taobao Ask, a service to enable users to obtain more-accurate product recommendations and search results on Taobao. Taobao Ask can answer questions via text, images, video, audio, and other content, which may help ensure users receive the most-relevant product recommendations.

Earlier this year, Tencent launched its own belated entry into the now highly competitive AI field. Its assistant app, Yuanbao, which is built on Hunyuan, its own LLM technology, can be used to analyse and summarise content, provide questions and answers, and generate text and images. Its biggest competitive advantage is its direct access to content within the app ecosystem on WeChat, Tencent’s super app, which has nearly 1.4 billion monthly active users. For instance, Yuanbao’s search results can pull from Official Accounts, the WeChat publishing platform where verified brands and creators post articles and information, which is not accessible to external search engines like Baidu or Google.

As generative AI–powered search gains momentum in APAC, users will increasingly spend time engaging in entire conversations and research queries with AI-driven chats, without the need to navigate multiple sites and articles. A single query could also result in related information and content that may well include suggestions and recommendations, making the consumer search journey quicker and more efficient.

Given this shift to a more intelligent future of search, how can marketers stay ahead, maximising the opportunities and minimising the risks? If anything, this evolution underscores the need for lots of quality content, be it text, images or videos. It will also be even more important to ensure that paid, owned, and earned content is integrated, as the AI-powered search engines work to process and consolidate everything about the brand, including advertising, reviews, and articles, into a single integrated experience.

Moreover, aggregating research and comparison information into a single source means that marketers will have to place greater emphasis on the authenticity of their content. To achieve that, brands need to better understand where they have a right to win, so as to build content authority and distinctiveness there, given that the AI algorithms in the search engines will present a form of opinion to the consumer.

As well, the danger of AI optimising content such that all results sound the same is very real, and therefore, ensuring a point of difference for the brand will be important. With generative AI here to stay, and as the premise of paid search evolves, marketers must pay even more attention to SEO and content optimisation to deliver an optimal brand experience to consumers.

Sharon Soh is the chief planning and audience officer for UM APAC.