Document Processing · Enterprise Knowledge Search
People's Daily needed to modernize news production for a large-scale mobile and multimedia audience while preserving editorial quality and safety. The organization had to handle huge volumes of video, image, and text assets, accelerate cloud-based editing, search, and review, and support AI-assisted content production across the full newsroom workflow.
Reported impact
People's Daily built an AI-assisted newsroom that can produce and review multimedia stories much faster than a purely local workflow. The system supports a client with more than 110 million downloads, can elastically create hundreds of cloud editing workstations during peak reporting periods, and gives editors faster search, automated asset generation, and safer content review for major reporting events.
People's Daily combined full cloud migration with an AI editorial platform from Alibaba Cloud. The client application moved onto Alibaba Cloud infrastructure using elastic architecture, CDN, video cloud, and managed operations so the media platform could scale for a large mobile audience. On top of that base, the Video Cloud AI Editorial Department became the newsroom automation layer. It processes video, image, and text assets for tasks such as cloud-based non-linear editing, intelligent poster generation, one-click portrait extraction, multimodal search, and intelligent review. Multimodal search combines face recognition, object recognition, speech recognition, OCR, video DNA, and natural-language processing so editors can find the right material quickly. Intelligent review uses AI plus big-data rules to flag content elements and scenarios before publication. Together, these systems give People's Daily an AI-assisted production line for faster multimedia journalism at newsroom scale.
Video, image, and text assets are uploaded into the cloud editorial environment, where AI editorial tools process them for editing, auto-generation, search, and review. Editors query the asset pool through multimodal retrieval, retrieve the needed materials, and publish optimized content back to the mobile client and broader media channels.
Multimodal search uses face recognition, object recognition, speech recognition, NLP, OCR, and video-DNA algorithms so editors can find assets more accurately and quickly. (original: 多模搜索功能借助人脸识别、物体识别、语音识别、自然语言处理、OCR 识别，及视频 DNA 的底层算法能力。)
People's Daily wanted to use cloud and internet technologies to serve a huge audience through a modern mainstream media platform. (original: 期望采用安全可靠、灵活弹性的云计算技术拥抱互联网，实现有新闻的地方就有《人民日报》、有用户的地方就有《人民日报》。)
The AI Editorial Department uses big data and AI to process massive video, image, and text assets and automatically generate video and graphic stories for specific themes. (original: 整个系统借助大数据和人工智能技术，对海量视频、图片、文字进行符合新闻业务标准的高效处理，针对设定好的主题，实现视频稿件、图文稿件的自动化、批量化、智能化生产。)
The upgraded system adds “look, understand, and review” capabilities, including intelligent poster generation, one-click portrait extraction, multimodal search, and content review. (original: 升级后的“人民日报 AI 智能编辑部”新增了多项功能，重点提升了“看”“悟”“审”能力。)
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
OpenAI Sample Mar 5, 2026 VfL Wolfsburg But modern football is no longer defined by performance on the pitch alone. Expectations from fans, partners, and internal stakeholders continue to rise—while budgets and headcount cannot scale indefinitely. ChatGPT Enterprise · ChatGPT Team · custom GPTs 
Alibaba Cloud Sample Jan 9, 2026 Bespin Global Bespin needed to industrialize enterprise AI agents without forcing customers to stitch together model stacks, orchestration frameworks, and multimodal data pipelines by hand. It also needed better document… delivery cycle reduced from 1 day to about 2 hours Alibaba Cloud Bailian · ChatOmni · CrewAI OpenAI Sample Aug 20, 2025 MIXI MIXI needed a secure company-wide AI environment so employees could share GPTs, use internal data safely, and speed up creative, investment, and communication work across a diverse consumer-services portfolio. 45-day deployment ChatGPT Enterprise · Copy Checker GPT · Creative Planning GPT 