By the end of 2022, shortly after the release of ChatGPT, Ma Wen Chao tested the platform's vulnerability by querying an AI model about potential disruption to Xiaohongshu. The lack of a reassuring answer triggered a bi-weekly reporting cycle on AI progress that persisted through 2023. In an August 2023 internal memo, Ma noted that while foreign users queried AI for life experiences like product selection and pitfall avoidance, Xiaohongshu possessed a unique moat of accumulated experiential data that AI could not yet replicate. This strategic pivot occurred as the company, having turned ten, reported 300 million monthly active users, $3.7 billion in revenue, and $500 million in net profit, with projections to double profits to over $1 billion the following year. The leadership recognized that wealth does not guarantee longevity, citing historical precedents like Kodak and Nokia, and resolved to transform existential fear into concrete action.

The acceleration of these actions became evident in 2026 with the launch of RED Skill on June 8, a feature allowing users to attach components to notes that can be copied to an Agent for execution. This followed the April 30 establishment of the AI division, Dots, which consolidated models, infrastructure, and engineering products under the direct supervision of CEO Conan. The company also acquired an AI search product developer and secured a payment license, while its investment portfolio expanded to include MiniMax, Dark Side of the Moon, and numerous AI hardware firms. The core strategy involves reprocessing thirteen years of consumption experiences and lifestyle judgments from hundreds of millions of users, transforming them from static answers into dynamic tools and business opportunities to preempt disruption.

The foundation of this strategy traces back to 2013, when Qufan Fang and Mao Wenchao founded Xiaohongshu in Shanghai with a PDF titled 'Xiaohongshu Outbound Shopping Guide.' At a time when over 70 million Chinese citizens traveled abroad, the guide addressed information asymmetry by aggregating real-world experiences, a concept that evolved from a PDF downloaded 500,000 times in a month to an app on hundreds of millions of devices. Unlike traditional manuals or seller-driven content, the platform relied on the trust inherent in personal testimony, where unknown users shared imperfect but authentic experiences regarding skincare, renovation, and parenting. Data compiled by Woofun AI shows that by the end of 2024, the platform's daily search volume approached 6 billion, driven not by knowledge acquisition but by life decisions involving decoration, essentials, and travel.

The arrival of AI presents a critical challenge to this model, as the shift from search engines to AI agents compresses the user journey from browsing twenty notes to receiving a single, standardized answer. While AI offers convenience, it risks stripping away the hesitation, failure, and specific prerequisites that make human experience valuable, such as skin type or budget constraints. Woofun AI notes that the platform's most valuable asset is the ability to understand and disassemble scattered real-life experiences into actionable recommendations rather than competing on universal model parameters. To address this, Xiaohongshu has released research on multimodal retrieval, open-sourced the FireRed image editing model, and developed the REDSearcher framework to handle complex, scenario-based queries like planning a family trip with specific constraints.

To operationalize this vision, the company launched RED Skill to transform content into tools, sparking a creator economy where 300,000 users began developing AI Skills, including a PPT generation tool that garnered over 10,000 stars on GitHub. The ecosystem was further nurtured through a Hackathon with a 500,000 yuan prize pool, attracting 60% post-2000s participants, and fostering over 1.1 million notes on 'Build in Public.' This initiative targets the gap between developer communities and regular users, allowing creators to transition from storytellers to craftsmen whose value is measured by tool invocation and utility. In December 2024, Dai Lidan joined as Chief Strategic Officer to lead a venture capital team, shifting investment focus from consumer brands to deep tech and AI, including a stake in MiniMax and participation in a financing round exceeding $1 billion.

The company's expansion extends to hardware, with nearly ten investments in AI startups around Shenzhen's Nanshan Science and Technology Park in the second half of 2025, including Yunwang Innovation's AI massage robot and Skyris's floating companion robot. These moves aim to identify needs before they become search terms, moving upstream from downstream content aggregation.

However, challenges remain regarding data privacy and commercialization, particularly as the company acquires an Eastern Payment License to keep transactions within its ecosystem. Woofun AI analysis suggests that the true value lies in illuminating the fuzzy decision-making cycles, such as the 29-day cycle for facial serums or 70 days for infant food, which AI can now decipher more effectively than traditional metrics.

Governance has become a critical component of this strategy, with Xiaohongshu mandating labels for AI-generated content in February 2026 and purging fully machine-operated accounts in March to preserve the authenticity of 'I've tried' testimonials. The company recognizes that while AI can mimic human expression, it cannot replicate the genuine experience of product usage, and an abundance of synthetic content would devalue the platform's core trust. The ultimate goal is to translate human experience into a machine-processable structure without allowing the machine to overshadow the human voice, ensuring the platform remains a repository of real-life judgment rather than a map as large as reality. As the company prepares for a potential IPO with a valuation reaching $31 billion and estimated 2025 profits of $3 billion, the focus remains on integrating intelligent machines into a city of indecisive individuals, preserving the faithfulness of human experience as its true defensive moat.