Supporters of Marcus Endicott’s Patreon can access weekly or monthly video consultations on this topic.
The report is titled "China AI Agent E-commerce Live Streaming Digital Human Real-World Assessment, 2Q25," published by IDC China in April 2025 under document number CHC53057125. It is classified as a Market Perspective, runs 14 pages with 9 figures, and was authored by senior IDC China AI analyst Wency Yang and analyst Len Li. It is priced at $20,000 and falls under IDC's China Automation and Agentic AI subscription series. It was the first report IDC had ever published specifically assessing this product category.
The report serves technology suppliers — vendors building or selling e-commerce digital human live streaming products — rather than enterprise buyers. Its geographic scope is China only. It does not assess the broader digital human market but focuses specifically on the e-commerce live streaming use case, where AI-generated virtual anchors conduct product sales broadcasts on behalf of merchants, operating 24 hours a day without human hosts. Seven vendors and their specific products were assessed:
SenseTime's Ruying, and
The report opens by mapping the overall development background of digital humans and then narrows to the current state of the e-commerce live streaming sub-sector. A dedicated figure presents a tiered development path for e-commerce live streaming digital humans, suggesting the report introduces a maturity framework or capability ladder that classifies vendors or products into developmental stages. The report identifies two dominant vendor archetypes that have emerged in the market: platform-type vendors, which leverage existing traffic pools and proprietary product data to build vertical-domain AI models optimised for commerce; and tool-enablement vendors, which focus on advancing visual rendering technology and building cross-platform solutions that can operate across multiple e-commerce ecosystems.
The report's core methodology is a hands-on real-world test rather than a purely analytical or survey-based evaluation. Vendors were assessed across five dimensions: technical capability, product performance, commercial maturity, compliance and platform stability, and customer service. Each vendor received an individual results figure, suggesting a scored or radar-chart style visualisation of performance across these five axes. The methodology section describes the testing approach in detail, and the assessment background section contextualises why these five dimensions were chosen as the evaluative standard.
Baidu's Huiboxing ranked first overall and led in four of the five assessment dimensions — technical capability, product performance, platform compliance and stability, and customer service — with commercial maturity apparently the dimension where the gap between vendors was less definitive. JD Cloud's Yanxi and Kuaishou's Nüwa placed second and third respectively, based on secondary media coverage. SenseTime's Ruying, Lingxi Shenzhi's Miaobao, Xiangxin Technology, and Jisi Technology placed in the remaining positions, though the report does not appear to publish an explicit numerical ranking beyond the overall winner.
Despite Baidu's leading position, the report identifies persistent weaknesses across the market as a whole. Emotional interaction remains homogenised — meaning most digital human anchors cannot yet deliver the kind of authentic, contextually responsive emotional engagement that distinguishes top human hosts. Content compliance risk is also flagged as an ongoing structural problem, as digital human anchors operating autonomously at scale create exposure to regulatory violations around advertising standards and platform content rules.
Lead analyst Wency Yang's quoted conclusion is notably cautious despite the commercial momentum the report documents. She frames the long-term viability of digital human live streaming around three unsolved challenges: technology must evolve from functional imitation of human hosts toward genuine emotional resonance; vendor ecosystems must balance platform dependency against the risk that dominant platforms could develop competing in-house solutions; and the regulatory environment must develop cross-platform standards to govern the sector. She explicitly states that many platforms remain cautious about scaling digital human anchors, and that whether this technology can achieve mass adoption at scale remains genuinely uncertain. Her conclusion is that sustainable competitive advantage will come from scenario-specific innovation and user experience design rather than raw technical capability, and that only vendors who achieve deep integration of technology, content, and commercial operations will unlock the full potential of digital humans as an e-commerce growth engine.
The report concludes with a dedicated section of recommendations addressed specifically to technology providers and service providers in the space, followed by pointers to further research and related IDC studies, and a full appendix. This structure suggests the report is designed to be actionable for vendors seeking to improve their competitive positioning, not merely descriptive of market conditions.
This is a rigorous, purpose-built competitive intelligence tool for vendors operating in or entering China's e-commerce digital human live streaming market. Its value lies primarily in the hands-on comparative testing methodology, the five-dimension scoring framework, and the individual vendor result figures, which together give any vendor a clear picture of where they stand against seven named competitors on specific capability axes. The analyst commentary adds strategic depth by tempering the commercial optimism surrounding the sector with a grounded view of the technical, regulatory, and platform-dependency risks that could constrain long-term growth.
Table of Contents
Executive Summary
Figure: China AI Agent E-commerce Live Streaming Digital Human Real-World Assessment, 2Q25
New Market Developments and Drivers
Overall Digital Human Development Background
Current State of E-commerce Live Streaming Digital Human Development
Figure: Tiered Development Path of E-commerce Live Streaming Digital Humans
Assessment Background, Criteria, and Methodology
Assessment Background
Five Core Indicators
Testing Methodology
Assessment Results
Vendor Overview
Baidu Huiboxing
Figure: Baidu E-commerce Live Streaming Digital Human Agent Assessment Results
Jisi Technology
Figure: Jisi E-commerce Live Streaming Digital Human Agent Assessment Results
JD Cloud Yanxi
Figure: JD Cloud E-commerce Live Streaming Digital Human Agent Assessment Results
Kuaishou Nüwa
Figure: Kuaishou E-commerce Live Streaming Digital Human Agent Assessment Results
Lingxi Shenzhi Miaobao
Figure: Lingxi Shenzhi E-commerce Live Streaming Digital Human Agent Assessment Results
SenseTime Ruying
Figure: SenseTime E-commerce Live Streaming Digital Human Agent Assessment Results
Xiangxin Technology
Figure: Xiangxin E-commerce Live Streaming Digital Human Agent Assessment Results
Recommendations for Technology and Service Providers
Further Research
Related Research
Appendix