New Methods, Organizations, and Tools: Huawei Unveils New Paths for AI Adoption in Industries

• Huawei unveiled a seven-step approach for AI adoption and the DIMAK engineering system. Backed by these new launches and its own industry-specific organizations, Huawei offers new methods, organizations, and tools to help industries adopt AI in core business scenarios and expand it at scale. This will enable three key transformations toward enterprise-wide intelligence.

SHANGHAI - At HUAWEI CONNECT 2026, Huawei unveiled a seven-step approach for AI adoption and the DIMAK engineering system. Backed by these new launches and its own industry-specific organizations, Huawei offers new methods, organizations, and tools to help industries adopt AI in core business scenarios and expand it at scale, according to the official website of Huawei. 

Tao Jingwen, Huawei's Deputy Chairman of the Supervisory Board and Chair of Global Industry Business Operations, delivered a keynote speech titled "Delving into Industry Scenarios to Bridge Digital and Intelligent Divides". He stated, "Currently, AI is advancing much faster than people's ability to adapt to it. AI has become more accessible, yet the challenge lies in how industries, employees, and organizations adapt to and harness it in the real world." According to Tao, tech firms need to not only explore and build powerful AI, but also address the challenges of how industries adapt to AI and how AI can understand industries.

He continued, "Future enterprise processes will be designed around collaboration between people and AI. This is not about mounting an engine onto an old carriage, but about designing a whole new car." To this end, new methods, organizations, and tools are needed to build "roads and bridges" between computing power and scenarios. This will allow AI to expand along the value stream in accordance with Gall's Law, upgrading single-point intelligence to system-level intelligence.

New methods: A seven-step approach that embeds AI into production systems

Tao discussed Huawei's seven-step AI adoption approach for resolving the challenge of deeply embedding agentic AI into enterprise systems. This approach starts small but goes deep. It starts from core scenarios and expands along the value stream, providing a replicable roadmap for the intelligent transformation of industries.

The seven steps are understanding an industry's business logic, selecting core industry scenarios, overcoming technical and engineering challenges, deploying AI in core production scenarios, expanding along the value stream, developing and collaborating with the industry ecosystem, and continually elevating industry intelligence. This approach has three key aspects.

• Selection of the right scenarios: Finding the right scenarios is half the battle. Business experts and AI tech experts need to work together and use Huawei's 12 Questions for Scenario Selection to identify high-value production scenarios along the value stream.
• In-depth AI adoption: Unlike experimental AI, which remains at the demonstration stage, the seven-step approach brings AI into real-world production and operations. It starts with developing an AI application as a minimum viable product (MVP), and then pilots it in specific scenarios, before integrating it into an entire enterprise production system. Finally, this application is deployed on the enterprise's IT infrastructure for routine operations. If the pilot is successful in one scenario, the application can be expanded along the value stream to create greater value.
• Long-term deployment: The next steps are integrating into the industry developer ecosystem and jointly developing best practices to promote the continued evolution of AI technologies and elevate industry intelligence.

"The seven-step approach provides a complete path for applying AI to core business scenarios, from understanding scenarios to embedding AI into production systems, from single-point intelligence to system-level intelligence, and from individuals to ecosystems," Tao concluded.

New organizations: Short-chain operations for closer partnerships with customers

Organizations need to be restructured to adapt to advances in technology. Last year, Huawei strengthened its own industry-specific organizations to further streamline research, marketing, sales, and service activities. The company now has 66 sub-business units (BUs) and account departments dedicated to specific verticals in 10 key industries. Experts from these organizations are engaged in frontline customer operations to ensure they have a thorough understanding of scenarios and can solve real-world problems. Furthermore, Huawei has established a computing platform team and industry-specific slim & agile teams. These teams are responsible for developing toolchains and engineering methods that convert complex technologies into plug-and-play services.

Tao stated, compared to previous long-chain operations, "short-chain operations allow us to quickly respond to customers' real needs". Huawei is committed to sharing with customers its digitalization experience spanning more than a decade and its cloud-computing-network-storage foundation, combined with the open models and technology ecosystems in the industry. "We not only provide tools and methods, but also seek to help customers build their own sustainable capabilities." Tao believes that in the era of AI, Huawei's relationships with customers will be upgraded from simply being buyers and sellers to partnerships centered on scenario-based joint research and innovation.

New tools: Three AI transformations enabled by the DIMAK engineering system

Huawei has released the Data, Infrastructure, Model, Agent, and Knowledge (DIMAK) engineering system. Built on practical experience, the engineering system provides technical capabilities for scenario-specific AI adoption in diverse industries.

First, the computing foundation is compatible with major models, enabling enterprises to use the models more quickly and efficiently while creating greater value. Second, data engineering and knowledge engineering are used to efficiently process core data assets and convert them into enterprise assets usable by AI. This integrates explicit knowledge such as enterprise regulations, policies, processes, and standards with the implicit knowledge of experts to form a business semantic foundation that can be understood and used by both people and AI. Finally, agent engineering helps enterprises design and govern agents for future needs based on business processes, specifying agents' authority, responsibilities, and how they collaborate with people.

"Discrete agents are like individual pearls—brilliant on their own, but incapable of forming a necklace without a unifying thread," explained Tao. "Unless orchestrated along business processes and adapted to business realities, they cannot deliver true value for enterprises. Worse, unorganized agents risk creating friction due to overlapping authority and conflicting responsibilities." He went on to explain that a collection of agents that are not organized around set processes ultimately cannot improve efficiency; instead, they may amplify workflow flaws and potentially trigger systemic breakdowns. Last year, to help address these challenges, Huawei open-sourced the openJiuwen AI Agent Platform, which supports multi-agent collaboration, end-to-end agent self-evolution, and enterprise-grade reliability and governance. "We will stay strictly within our boundaries, not overstepping our role in the industry, while delivering the technical tools needed to bridge the gap in AI adoption in industry scenarios."

Currently, AI is moving from auxiliary to central enterprise operations. Systematic architecture and engineering support will be needed to unleash the value of AI at scale. Tao discussed how Huawei will work closely with industry customers and use industry-specific organizations and the DIMAK engineering system to drive three key AI transformations across industries:

• Translating general-purpose intelligence into enterprise-specific intelligence to allow AI to truly understand business;
• Translating enterprise-specific intelligence into trustworthy actions to ensure that AI actions are controllable and reliable in governance;
• Translating scattered intelligence into enterprise-wide intelligence to enable collaboration for continual value creation based on enterprise goals.

The DIMAK engineering system has already been implemented in numerous industries. At the event, executives from Shenzhen Loop Area Institute, Nanjing Iron & Steel Co., Ltd., and Guangzhou Phar. Holdings shared their success stories in operator development and model adaptation and tuning, data and knowledge governance in the steel and logistics industries, and accelerated drug R&D with AI.

Huawei also released the Agentic Enterprise white paper and a collection of 157 practical case studies of digital and intelligent transformation, which serve as references for the intelligent upgrades of global industries.