Recently, in the "AI Leading the Tide · Professional Geek Competition" track of the Beijing AI Innovation Competition, Sointech (SOIN AI) won the special award of the "Scenario Application Professional Geek Competition" with its entry "Paperlight: Next-Generation AI Video Generation Infrastructure." The award focuses on the application value of AI technology in real business scenarios, reflecting SOIN AI's continued exploration and practical achievements in multimodal agent applications.
The Beijing AI Innovation Competition is an important municipal AI competition, jointly organized by the Cyberspace Administration of the CPC Beijing Municipal Committee, the Mentougou District People's Government and other units. Centered on "promoting innovation and application through competition," it selects innovative achievements that promote deep integration of AI and the real economy. SOIN AI's award in the "AI Leading the Tide · Professional Geek Competition" track marks that its capabilities in enterprise AI applications and multimodal scenario deployment have been recognized by the municipal innovation system.
Driving Intelligentization of Complex Business Processes with Agentic AI
SOIN AI, a subsidiary of Sointech (stock code: 871539), is a high-tech company focused on enterprise-grade Agentic AI applications. Around the goal of "letting AI truly enter core enterprise processes," the company has built a product system centered on agent workflows, serving advanced manufacturing and highly complex document-intensive industries.
SOIN AI has formed a product architecture with the Agentic Workflow Engine (AWE) as the foundation, working in concert with the Agentic Documents Suite (ADS) and the Autonomous Lifecycle Assistant (ALA). Rather than providing isolated AI features, this system embeds AI capabilities into enterprises' existing systems and business processes through agent collaboration and process orchestration, supporting gradual intelligent upgrading while enterprises run stably.
The Award-Winning Project Paperlight: "Process-Type Innovation" in Multimodal Content Production
The award-winning project "Paperlight: Next-Generation AI Video Generation Infrastructure" originates from SOIN AI's long-term practice in education publishing and enterprise content production. The project focuses on coordinated production across multimodal content — text, images, voice and video — restructuring the complete process from content understanding to final video output through agents, exploring a scalable deployment path for AI video in professional scenarios.
In practical application, Paperlight targets high information-density content such as papers, white papers and research reports, providing intelligent support from structure understanding and content reorganization to video expression, helping institutions improve content distribution efficiency and expression diversity while maintaining professionalism. The award also reflects the evaluation system's positive view of the direction of "AI as infrastructure for production processes."
The Product System Supporting Multimodal Capabilities
SOIN AI's award in the fiercely competitive AI innovation track is inseparable from its years of technical accumulation in enterprise-grade Agentic AI. The company's core product system — the Agentic Workflow Engine (AWE), the Agentic Documents Suite (ADS) enterprise intelligent document system, and the Autonomous Lifecycle Assistant (ALA) — has been validated in highly complex vertical fields such as advanced manufacturing, finance, and education publishing.
In key technical dimensions such as multimodal document understanding, intelligent content generation and automated workflow orchestration, SOIN AI has established industry-grade technical barriers. Especially in end-to-end solutions of "deep document understanding + AI agent workflows," SOIN AI can precisely meet enterprises' stringent requirements for content accuracy, style consistency and process controllability.
Paperlight's practice is not isolated product innovation, but a concentrated demonstration of SOIN AI's overall product system capabilities:
AWE (agent workflow engine), as the underlying capability, supports multi-agent collaboration, task orchestration and tool scheduling, making complex content production processes configurable and evolvable;
ADS (enterprise intelligent document system) has accumulated capabilities in multimodal document understanding, content generation and process automation, providing the foundation for high-quality content input and structuring;
ALA (R&D lifecycle intelligent assistant) has verified in manufacturing and other complex scenarios the feasibility of "embedded agents" working in concert with enterprise systems, providing a methodology for reliable AI deployment in serious business.
Through this system, SOIN AI continues to explore how to advance AI from the "tool layer" to the "process layer" and "execution layer," letting agents truly participate in daily enterprise operations.
Taking the Competition as an Opportunity, Continuously Deepening Industry Scenario Deployment
Winning at the Beijing AI Innovation Competition is an important affirmation of the company's technical route and product direction. Going forward, SOIN AI will continue to focus on core enterprise business scenarios, deepen Agentic AI deployment in document intelligence, R&D management and multimodal content production, and steadily advance product application expansion across industries.
Against the backdrop of deepening industrial intelligentization, SOIN AI will adhere to the product philosophy of "embedded, scenario-based, sustainably evolving," joining hands with industry partners to explore the long-term value path of AI empowering the real economy.