My honest take on Xiaohongshu's AI shopping assistant after testing it

Title: Xiaohongshu AI Shopping Guide, My Honest Thoughts After Trying It

As someone who frequently tinkers with various AI tools, I tried Xiaohongshu's AI shopping guide. Specifically, I tested the already-launched "Diandian," the official intelligent search assistant. News reports say Xiaohongshu is promoting a new AI shopping guide led by Daoxuan, but currently, the only forms I can access are Diandian and AI summaries in store reviews. So this review focuses mainly on the Diandian experience, with some thoughts on potential directions for the new feature.

First, let me explain how to find it. I clicked the AI entry directly in the Xiaohongshu search box, or you can @mention Diandian on the conversation page. My first question was "What bag should a beginner commuter buy?" It replied in about three seconds with a fairly complete list, ranging from tote bags to underarm bags, with note links attached below each option. Honestly, this experience was a bit surprising. Unlike traditional searches that give you a pile of notes to filter yourself, it gives you conclusions directly and lets you click in to see the basis. I checked, and the recommendation logic indeed leverages Xiaohongshu's content library, reportedly having processed 1 billion notes, so the posts it cited matched well in terms of time, scenario, and price range.

But pitfalls came quickly too. When I asked my second question, "Phone recommendations under 4,000 yuan," it gave several options. When I clicked on the product cards, I saw two were brand-sponsored notes ranking high, with comments showing signs of "paid promotion." I can't say the recommendations were wrong, but there's an issue: it mixes advertising content with genuine user shares. As a shopping guide, the boundary between neutral recommendations and commercial partnerships is very blurry. Based on my testing, it feels more like "helping you organize notes into answers" rather than "helping you make purchasing decisions."

Comparing with other players: Taobao relies on Tongyi Qianwen for end-to-end guidance from recommendation to order; Doubao directly compares prices and calculates discounts; ChatGPT also launched shopping features last year, providing illustrated recommendations based on needs. Xiaohongshu's differentiation lies in having the strongest pool of "seeding" (product recommendation) content, but its e-commerce transaction process has always been less complete than Taobao's. So the key for this new AI shopping guide isn't whether recommendations are accurate, but whether it can keep transactions on-platform after mounting order redirects. If it just directs traffic to brands, it remains an ad slot, not a shopping guide.

From a strategic perspective, Xiaohongshu established the first-tier AI department Dots in late April, and now Daoxuan is personally pushing the shopping guide project, indicating that e-commerce is the top priority for AI implementation. This direction is correct, after all, what makes Xiaohongshu most valuable is the scenario of "users searching with purchase intent." But the core competitive moat isn't AI technology, but community trust. If the AI shopping guide becomes a soft-ad distributor, and users realize after a few uses that recommended brands are those who paid, the feature will fail. Conversely, if it can clearly label which are ads and which are genuine reviews, and smooth out the ordering link, it might be the closest to "conversion" among all AI shopping guides.