Huawei’s Three Key Steps To Drive Industrial Intelligence Forward And Unleash Ai’s Full Potential

Leo Chen, Huawei’s Senior Vice President And President Of Enterprise Sales. Credit: Huawei

Speaking astatine HUAWEI CONNECT 2025 successful Shanghai, Leo Chen, Senior Vice President and President of Enterprise Sales astatine Huawei, reminded nan assemblage that AI is advancing faster than ever. “No 1 is asking why we should usage AI,” he said. “The existent mobility is how.”

Taking to nan shape for his keynote reside Breaking New Ground to Industrial Intelligence, Chen emphasised that AI is present to enactment and will importantly boost business efficiency. However, he says, location are 3 awesome questions facing enterprises trying to adopt AI: really to guarantee AI investments present business value, really to leverage proprietary information for competitory advantage, and really to standard AI usage cases beyond pilots. 

Discover B2B Marketing That Performs

Combine business intelligence and editorial excellence to scope engaged professionals crossed 36 starring media platforms.

Find retired much

Chen introduced Huawei’s latest attack to business intelligence, nan three-step “ACT” pathway – Assess, Calibrate, Transform – designed to guideline enterprises done intelligent transformation. This pathway helps enterprises place high-value AI scenarios, calibrate models pinch vertical data, and toggle shape operations pinch scalable AI agents. Huawei besides launched 9 awesome solutions for business intelligence that it has developed alongside its partners.

“Real knowledge comes from practice,” says Chen, who shared occurrence stories from respective different industries:

Case study 1: Intelligent powerfulness statement inspections pinch AI

China Southern Power Grid oversees extended powerfulness transmission infrastructure, generating immense amounts of operational data, and wanted to leverage this information to amended nan recognition of defects successful powerfulness lines. The superior situation was to efficiently process and analyse ample volumes of information to observe defects successful powerfulness transmission lines, aiming for higher accuracy and automation.

In collaboration pinch Huawei, China Southern Power Grid developed “MegaWatt” – a ample AI exemplary tailored for nan powerfulness sector. Built connected Ascend computing level and MindSpore AI framework, MegaWatt integrates machine imagination and earthy connection processing to automate defect detection. The task progressive broad information governance, including cleaning, processing, labelling and optimising endeavor data. Huawei’s optimised operators accelerated exemplary training and improved accuracy.

Operating connected Ascend’s MoE Expert Parallelism (EP) cluster, designed for high-volume, real-time tasks, this setup accrued per-card throughput by 3.3 times compared to accepted solutions and importantly reduced conclusion latency. Since deployment, MegaWatt has accrued defect nickname ratio by 5 times, improved accuracy to complete 90%, and enabled end-to-end automation from image nickname to study generation.

Case study 2: Transforming aesculapian believe pinch AI

Accurate and businesslike aesculapian recordkeeping is basal for value diligent care. West China Hospital sought to streamline this process to amended some objective outcomes and unit ratio and required a solution that could make high-quality records quickly and accurately, while gathering strict objective standards.

In business pinch Runda, Huawei developed an AI-powered aesculapian grounds strategy based connected nan Ascend platform, integrating open-source models for wide tasks pinch industry-specific models for objective connection understanding. The strategy automatically recognises and summarises doctor–patient conversations, interprets diagnoses and generates meticulous infirmary records.

Deployed astatine West China Hospital, nan strategy useful alongside pre-diagnosis and value power AI agents to guarantee some ratio and accuracy. Doctors tin finalise records pinch nary much than 4 edits and taxable them to nan infirmary accusation strategy (HIS) pinch a azygous click. The AI-powered strategy generates aesculapian records successful conscionable 1 second, importantly improving nan ratio of test and treatment.

Five cardinal findings for driving business intelligence

From these and different projects, Chen explained that Huawei had identified 5 cardinal findings for successful intelligent transformation:

1. Selecting nan correct implementation scenarios is important for realising AI’s value.
2. High-quality, industry-specific information is basal for effective AI models.
3. The request for large-scale conclusion is quickly increasing.
4. Human-AI collaboration is emerging arsenic a caller organisational model.
5. Strong governance and consequence guidance are captious for secure, sustainable AI.

Huawei is supporting enterprises utilizing nan ACT pathway and helping them measure high-value AI scenarios, says Chen, while Huawei’s afloat toolchain and information systems alteration nan creation of industry-specific models. Its Versatile level accelerates supplier deployment and workflow automation, while nan institution besides invests successful AI-oriented ICT infrastructure, specified arsenic high-speed networks and precocious retention solutions, to support nan full AI lifecycle.

Additionally, Huawei is building a robust partner ecosystem, now including complete 6,300 Kunpeng partners and 2,700 Ascend partners. Huawei’s 9 caller solutions – launched astatine nan Shanghai arena – screen areas for illustration metropolis management, intelligent computing, healthcare, banking, manufacturing, logistics, energy, and alloy accumulation and purpose to thrust nan adjacent shape of business intelligence.

Chen concluded pinch this thought: “Our corporate consequence to nan changes brought by AI will find whether it tin genuinely present its “last mile” worth and toggle shape exertion breakthroughs into real-world benefits.”

For much information, sojourn https://e.huawei.com/en/