Zhongke Wenge's Decision AI Product System at WAIC 2026

Zhongke Wenge frames its new product line as a single organism rather than six separate tools. At a forum during WAIC 2026 in Shanghai, CEO Luo Yin described a 'Decision AI Tree' (决策AI树) rooted in real business operations^[1]: 海汇TokSea sits at the roots as enterprise token infrastructure handling model access, routing and cost monitoring; the DIP platform is the hub, turning fragmented enterprise data into a business 'digital twin'; YaYi (general-purpose) and ScienceOne (co-developed with the Chinese Academy of Sciences for scientific literature and multimodal analysis) form the core intelligence engines; Decitron, billed as the 'brain,' combines world models, multi-agent simulation and game-theoretic solving to model problems and simulate outcomes; and Claworks, the canopy, is an agent platform that executes tasks directly through WeChat, DingTalk or Feishu. The whole stack runs on what the company calls a DOMA (Data/Ontology/Model/Agents) architecture^[2]. Luo Yin's framing is explicit: 'AI's next stage must move beyond content generation into deep participation in enterprise decision-making'^[3], and elsewhere that general LLMs 'can only answer the present, while Decitron calculates the future'^[4].

The product suite lands squarely inside Zhongke Wenge's IPO narrative. The company - founded in 2017 by scientists from the Chinese Academy of Sciences' Institute of Automation^[5]- cleared its Hong Kong Stock Exchange main-board listing hearing on June 9, 2026^[6], priced at a roughly RMB 4.046 billion valuation after ten funding rounds^[7], and debuted on June 26, 2026, with shares surging more than 81% on the first trading day^[8]. It launched Decitron, its flagship 'general-purpose decision-making LLM,' in Beijing on June 5, ahead of the full WAIC reveal^[9], then TokSea as a standalone access layer on June 30^[10], before assembling everything into the Decision AI Tree narrative at WAIC. The company reports 2025 revenue of RMB 405.3 million, up from RMB 317.8 million in 2024 and RMB 249.7 million in 2023, and claims the top market-share position in China's enterprise LLM-driven decision-intelligence segment^[11]. Existing customers cited include China Sinopec, China Merchants Bank, Bank of Beijing, Haier, BYD, Xinhua News Agency, People's Daily, CCTV and China's Ministry of Public Security, alongside a reported drop in average deployment cycle from 185 days in 2023 to 80.2 days and a 139.5% net revenue retention rate for 2025^[12]. Sequencing a full product-suite launch this close to a blockbuster IPO reads as much as investor-relations messaging as it does a product announcement.

Zhongke Wenge backs the launch with striking numbers, though the sourcing is uniformly company-reported and repeats across near-identical wire pickups of the same press event rather than independent benchmarks. CEO Luo Yin is credited with claims that the product system cuts enterprise AI construction and operating costs by more than 90%^[13][14]and improves decision-response speed by up to 192x^[15]. The company also points to a fast-growing addressable market, citing Zhize Consulting projections that China's enterprise decision-AI market will grow roughly tenfold, from RMB 39 billion in 2025 to RMB 375 billion by 2030^[13]. Decitron specifically is said to have produced a 160.28% annualized return in a 28-day simulated trading test based on its probability signals^[15]- a backtest result, not a live-market track record. Notably, coverage of the same 'days-to-minutes' compression claim is internally inconsistent across outlets covering the identical event - reported baselines vary between roughly '2-3 days' and '32 hours' for workflows now said to take about 10 minutes - a discrepancy that suggests rounded marketing figures rather than a single audited benchmark, and a reason to treat the headline percentages as directional rather than precise.

Public reaction to the WAIC unveiling itself appears limited so far, though Hong Kong-focused investor and finance channels have separately been framing Zhongke Wenge - around its recent listing - as a state-aligned 'AI decision brain.' That framing centers on investment narrative rather than technical reaction to the Decision AI Tree product suite itself, and engagement around it remains modest rather than indicative of broad community buzz.