Supporters of Marcus Endicott’s Patreon can access weekly or monthly video consultations on this topic.
The IDC PeerScape: China AI Digital Human Construction Best Practices, 2025 is a six-page practitioner-oriented research report published in October 2025 by IDC China, authored by AI Research Manager Anne Cheng (程荫). It is priced at $15,000 and sits within IDC's China AI and Generative AI Strategies subscription program. As a PeerScape format — as opposed to a MarketScape, which ranks vendors — its purpose is documentary rather than evaluative. It does not score or position companies against each other. Instead, it extracts real-world implementation patterns from enterprise deployments and distills them into transferable guidance for technology buyers considering similar investments.
The report addresses a specific gap in the market at a specific moment in time. By late 2025, AI digital human technology in China had moved past early experimentation and into genuine enterprise deployment, but enterprises — particularly those without deep AI expertise — were struggling to understand how to implement it well. IDC's value here is to compress the learning curve: by documenting what worked, what the challenges were, and what guidance can be drawn from real deployments, the report gives procurement decision-makers and CIOs a framework for thinking about their own investments. The $15,000 price tag signals that the primary buyer is an enterprise IT or strategy team, not an individual researcher.
The report is organized around four discrete best practices, each built on the same three-part structure of challenge, example, and guidance. This architecture is deliberate and practical. It acknowledges upfront that implementation is hard — each practice begins with an honest account of what makes that dimension of digital human deployment difficult — and then shows how a real organization navigated those difficulties before drawing generalizable lessons. The four practices move from the technical to the strategic: the first two address how to build and deploy digital humans effectively, while the latter two address how to deploy them broadly and how to think about their long-term value to the organization.
The first practice — full-stack technology integration of AI digital humans with data, large models, and AI agents — represents the most technically demanding and forward-looking recommendation in the report. The case study is a central state-owned enterprise whose implementation was delivered by Zhongshu Ruizhi (中数睿智), a specialist AI agent infrastructure company. The anonymization of the enterprise client is significant: it reflects both client preference for confidentiality and the politically sensitive nature of state enterprise AI procurement in China. IDC's key challenge finding here is that most enterprise buyers underestimate the complexity of deploying agent and large model technologies together, and that successful implementation requires a vendor with not just strong technology but a proven track record of customized, iterative delivery in comparable enterprise environments.
The second practice — democratization of AI digital human technology — signals a maturation of the market. The case study is New Oriental International Education, one of China's largest private education groups, and its inclusion is telling. New Oriental is not a technology company and does not have deep AI infrastructure expertise, yet it is deploying AI digital humans at scale. IDC's argument here is that the falling cost and complexity of digital human development tools is enabling organizations in education, retail, and other non-tech sectors to become serious adopters. The guidance derived from this practice is oriented toward helping enterprises identify what pre-built capabilities exist and how to avoid over-engineering solutions when simpler approaches are now viable.
The third practice — driving deployment across multiple use cases and scenarios — addresses a trap that many early adopters fall into: deploying a digital human successfully in one context and then failing to scale it across the organization. IDC frames this as both a technology challenge and an organizational one, requiring enterprises to think systematically about where digital humans create value rather than treating each deployment as a standalone project.
The fourth practice is the most strategically oriented and arguably the most important for senior leadership audiences. It argues that the long-term business case for AI digital humans goes well beyond cost reduction and efficiency metrics and extends to brand differentiation, customer experience, revenue growth, and competitive positioning. This reframing — from AI digital humans as a cost tool to AI digital humans as a value-creation platform — is consistent with IDC's broader messaging about the shift from AI experimentation to AI-driven enterprise reinvention.
Three companies are named as covered in the report: Huawei, Bank of Communications (交通银行), and SenseTime. Their selection is not arbitrary. Huawei represents the technology infrastructure layer — as both a vendor and an enterprise deployer of AI at scale, it anchors the report's technical credibility. Bank of Communications represents the financial services sector, which is by far the largest enterprise buyer of AI digital humans in China's B2B market, and its inclusion signals that the report speaks directly to the financial industry's specific deployment context. SenseTime, one of China's foremost AI companies and a major player in visual AI and digital human technology, lends the report authority on the vendor side and signals that the practices documented here reflect capabilities that are commercially available and production-tested.
The report was published against a market that IDC itself tracks closely. China's AI digital human market reached 4.12 billion RMB in 2024, growing 85.3% year-on-year, and IDC projects it will reach 12.5 billion RMB by 2027. The 2D digital human segment is currently the dominant category due to lower production costs and faster deployment timelines, though 3D digital humans are expected to grow as generative AI rendering technology matures. The technology stack underpinning the market is also shifting structurally: traditional production components like motion capture, graphic rendering, and image library construction are being progressively replaced by AI-native generation, which is shortening the production chain and further reducing costs.
Because it is a PeerScape and not a MarketScape, this document carries a different kind of authority. It does not tell you which digital human vendor to buy. What it tells you is how leading enterprises in China are approaching deployment, what mistakes are common, and what practices correlate with successful outcomes. For a technology buyer evaluating whether and how to invest in AI digital humans, it functions as a strategic primer grounded in real implementation experience rather than vendor marketing. For a vendor trying to understand what enterprise clients care about, it maps the decision criteria and concerns of sophisticated buyers across multiple industries. For an investor or analyst trying to understand where the China digital human market is headed, it provides a ground-level view of how the technology is actually being absorbed into enterprise operations — which is ultimately the most reliable signal of where commercial value is concentrating.
The report's most obvious limitation is its length. At six pages, it is compressed to the point where each of the four practices — each of which could sustain a much deeper treatment — receives only a few hundred words. The anonymization of the central state-owned enterprise in Practice 1 reduces transparency around one of the report's most technically ambitious case studies. And because IDC's PeerScape reports are built from cases that vendors and clients agree to share, there is an inherent selection bias toward success stories. The challenges documented are real, but they are challenges that were overcome — the report does not capture failed implementations or abandoned projects, which would provide equally valuable guidance.
Despite these constraints, it remains the most authoritative structured documentation of enterprise AI digital human deployment practice in China available at the time of its publication, and Anne Cheng's positioning as IDC China's lead AI analyst gives the interpretive framework that holds it together genuine market credibility.
Table of Contents:
IDC PeerScape Figure
Figure: IDC PeerScape: China AI Digital Human Construction Best Practices, 2025
Executive Summary
Peer Insights
Practice 1: Full-Stack Technology Application — AI Digital Human + Data + Large Models + AI Agents
Challenges
Example
A Central State-Owned Enterprise
Guidance
Practice 2: Democratization of AI Digital Human Technology
Challenges
Example
New Oriental International Education
Guidance
Practice 3: Driving Multi-Dimensional Scenario Applications
Challenges
Example
Guidance
Practice 4: Long-Term Value Lies in Enterprise Brand Building, Competitiveness, Revenue Growth, Customer Base Expansion, and Cost Reduction & Efficiency Improvement
Challenges
Example
Guidance