Promoting AI application in production and business operations

Notably, the three-step roadmap “ACT” (Assess – Calibrate – Transform) aims to promote the application of artificial intelligence (AI) in production and business operations. At the same time, he also introduced 9 smart industrial solutions developed with partners in the global ecosystem, aiming to realize large-scale digital transformation in the industry.

Mr. Leo Chen said that the rapid development of AI in recent years is posing three major challenges for businesses: how to make AI investments bring real business value; how to exploit internal data to maintain competitive advantage; and how to make AI applications pass the testing stage to large-scale deployment.

Mr. Leo Chen – Senior Vice President and President of Huawei Enterprise Business Group.

Mr. Leo Chen shared how Huawei solves these problems through three successful business lessons corresponding to three typical industries: banking, electricity and healthcare .

In the banking industry, Huawei develops a flexible master-slave agent architecture that integrates system engineering capabilities into business processes. This solution has been applied in various scenarios such as transactions, customer interactions, and risk control, helping banks increase revenue, improve operational efficiency, and manage risks more effectively.

In the power industry, China Southern Power Grid has applied Huawei's Ascend computing platform and MindSpore AI framework to develop a large-scale “MegaWatt” model that combines computer vision and natural language processing. The solution has helped customers improve the efficiency of risk identification during smart power line inspection by 5 times, while increasing the accuracy of image recognition to over 90%.

In the medical industry, Runda Medical used Huawei's Ascend server to develop an AI medical record creation solution, which shortened the record creation time to about one second, significantly improving the efficiency of medical examination and treatment at West China Hospital.

From the actual projects implemented, Mr. Leo Chen draws out five key factors that promote the intellectual transformation of the industry.

First, choosing the right application scenario is a key success factor, because AI only creates value when deeply integrated into core production scenarios.

Second, specialized AI models need to be trained from high-quality data, instead of using generic models that are ineffective in industrial environments.

Third, the need for large-scale inference will continue to grow rapidly with the development of AI agents, further driving the need for large-scale inference.

Fourth, human-AI collaboration is emerging as a new organizational model.

Fifth, system governance and risk control are increasingly important to ensure AI operates safely, sustainably, and reliably.

At the event, Huawei representatives also proposed a three-step “ACT” roadmap including: Assess - evaluate high-value scenarios, Calibrate - calibrate AI models using specialized data, and Transform - transform business operations with large-scale AI agents.

Huawei uses the AI Scenario Assessment Framework to evaluate high-value scenarios based on three factors: business value, scenario maturity, and technology-business integration. This framework has helped customers identify more than 1,000 core production scenarios that can be applied to AI.

Huawei also provides a comprehensive toolkit and AI security system to help enterprises calibrate general models and build industry-specific models. The toolkit uses vertical data to convert enterprise raw data into knowledge, and knowledge into models.

In the final step of the ACT roadmap – transforming businesses with AI agents at scale, Huawei introduced the “one-stop” Versatile platform, which can automatically generate agents and processes with more than 100 steps. This platform helps speed up the deployment of AI agents. In addition, Huawei has developed an AI talent cultivation program, supporting business professionals to be effective in developing, deploying, and operating AI agents.

Enterprises looking to deploy along the ACT roadmap will need an AI-oriented ICT infrastructure that covers the entire process from data preparation and processing, to model training, inference, and application development.

Nhat Linh

Source: https://doanhnghiepvn.vn/kinh-te/thuc-day-ung-dung-ai-trong-san-xuat-van-hanh-doanh-nghiep/20250925090401159