Hotel and Airline Brands Are Nearly Invisible in China's AI Travel Boom in 2026

Doubao passed 300 million monthly active users in Q1 2026, and it is increasingly the first stop for Chinese travelers planning a trip abroad — not just for inspiration, but for itinerary drafts that plug directly into hotel and flight inventory. Yet when we tested queries like "日本亲子温泉酒店推荐" (family-friendly onsen hotel recommendations, Japan) and "东南亚高性价比航班推荐" (best-value Southeast Asia flights) across Doubao, DeepSeek, and Qwen, the answers were dominated by OTA aggregators and a handful of Chinese-market-savvy chains — not the global hotel and airline brands with the strongest actual product fit. This is a category hubGEO has not covered before, and the gap is stark enough to matter for any hospitality or travel brand with China ambitions.

Why travel queries are different from other categories

Most GEO analysis on this site has focused on categories where AI models answer a discrete "which brand" question — skincare, sneakers, cars. Travel queries are structurally different: they are multi-step, transactional, and increasingly resolved without the user ever typing a brand name. A model doesn't just need to know that the Peninsula or Marriott exists — it needs structured, current inventory, pricing, and availability to complete the task the user actually asked for.

According to reporting reviewed for this piece, one commercial AI-travel workflow now combines DeepSeek's data-processing strength with Doubao's natural-language understanding to generate a full Xiaohongshu-style itinerary — dates, budget, and group size in; hotel inventory and real-time pricing synced out — in about three minutes. That is the workflow international hotel and airline brands are competing inside of, and most are not present in it at all, because their Chinese-market inventory and content sit outside the OTA integrations (Ctrip/Trip.com, Fliggy, Meituan) these assistants are wired into.

This matters because it changes what "visibility" even means for the category. For a skincare or sportswear brand, the win condition is being named in a recommendation. For a hotel or airline, the win condition is being represented accurately and completely in a structured feed the model can query — a much higher, and much less content-driven, bar.

The platform map has just been redrawn

Three months of MAU data make the platform picture unusually volatile right now, and travel brands need to recalibrate before locking in a GEO budget for the second half of the year:

Figures are drawn from multiple March–April 2026 industry reports (QuestMobile, 36Kr, and trade press) and vary by methodology, but the direction is consistent across sources: Doubao and Qwen are consolidating consumer travel-planning traffic, DeepSeek is becoming the default for comparison-style research, and Kimi's collapse means the long-context itinerary-parsing use case travel brands built content for over the past year now reaches a much smaller audience than it did in late 2025.

Where hotel and airline brands are actually losing visibility

Three patterns showed up repeatedly across our test queries, and they compound each other.

OTA aggregators outrank direct brand content. When a model can pull live inventory from an integrated OTA, it will generally prefer that structured data over parsing a hotel's own English- or lightly-localized Chinese-language site. In our tests, answers for "曼谷五星酒店推荐" (five-star Bangkok hotel recommendations) surfaced Ctrip and Fliggy package listings first, with individual hotel brand names appearing only as line items inside those listings — not as the source the model reasoned from. Brands without clean, current listings inside Ctrip, Fliggy, or Meituan are effectively invisible to the "advice-to-booking" workflow, regardless of global brand strength.

Video and UGC content shapes Doubao's answers more than owned content does. Because Doubao's search leans on multimodal and Douyin-sourced video, hotel and airline brands with little short-video presence in the Chinese ecosystem — as opposed to polished English-language brand films hosted on a global site — are simply absent from a large share of Doubao's travel answers. A brand can have an award-winning international marketing campaign and still not exist, from Doubao's perspective, if that content was never localized into short-form Chinese video.

Category language doesn't translate. A "boutique hotel" or "premium economy" framing that works for a Western audience often doesn't map cleanly onto the categories Chinese travelers actually query for — 亲子 (family-friendly), 蜜月 (honeymoon), 商务 (business travel), 性价比 (value-for-money). Models cite whichever source most clearly answers the query as asked, in the traveler's own category language, and most international hospitality content simply isn't written that way. This is the same "entity clarity" problem hubGEO has documented in luxury and beauty categories, but travel adds the further complication that the citation itself is often a booking widget rather than an article.

Airlines face a narrower but sharper version of the same problem

Airlines see a tighter version of the hotel dynamic. Flight search is already dominated by fare-aggregation logic — lowest price, fewest stops, best schedule — so an airline's brand rarely gets a chance to differentiate on service, lounge access, or loyalty benefits inside an AI answer unless the traveler asks a follow-up question specifically about those things. In our tests, queries like "东京到首尔哪个航空公司比较好" (which airline is better, Tokyo to Seoul) returned fare comparisons pulled from aggregator data, with carrier reputation mentioned only in passing, if at all.

The exception is where an airline has built distinct, citation-ready Chinese content around a specific traveler concern — baggage policy for skiing equipment, pet-in-cabin rules, or transfer times at a particular hub. Those narrow, fact-dense pages are exactly the kind of source models can quote directly, and they showed up far more often in our tests than general brand or loyalty-program marketing pages did. This is a smaller, cheaper fix than the OTA-integration problem hotels face, and it's one most airline marketing teams could execute this quarter without new vendor relationships.

What changed since hubGEO's last China AI-search snapshot

Prior coverage on this site — including Kimi: Second to Eighth and the China AI MAU Rankings piece — tracked the broader consumer and enterprise reshuffling across categories in general terms. Travel adds a wrinkle those pieces didn't need to address: because travel queries are increasingly transactional rather than purely informational, the relevant "citation" for an AI answer often isn't a blog post or press release at all — it's a live inventory feed. That means the standard GEO playbook (answer-first pages, entity clarity, citation-ready fact density) is necessary but not sufficient for hospitality and airline brands. OTA data hygiene now matters just as much as content strategy, and it's a layer most global brand marketing teams don't own or even monitor.

It's also worth noting what these numbers do not show. None of the reports reviewed here break out travel-specific query volume or travel-brand citation rates the way hubGEO's brand tracker does for categories like beauty or autos — this is genuinely underbuilt data, and the estimates above should be read as directional rather than precise.

Takeaway for Brand Marketers

If you market a hotel group, airline, or travel brand into China, three things are worth doing this quarter. First, audit how your inventory actually appears inside Ctrip/Trip.com, Fliggy, and Meituan — these feeds are increasingly what AI travel answers are built on, not your own site, and a stale or incomplete listing there is now a bigger visibility risk than a weak blog. Second, rebalance content investment away from Kimi-specific long-context formatting (multi-day PDF itineraries) and toward Doubao and DeepSeek, given the audience shift documented above; a format built for a shrinking sub-10M-MAU platform is a much smaller bet than it was six months ago. Third, build short-video and UGC-style Chinese content around specific, query-shaped use cases (family travel, honeymoon, business trip) rather than relying on translated brand marketing — Doubao's multimodal search rewards exactly this kind of content, and it's a gap almost no international hospitality brand has closed yet.

Related: see hubGEO's brand visibility tracking for how individual brands score across China's six major AI models, and our prior coverage of the Doubao/Qwen/DeepSeek MAU realignment.