经济新闻
新举措助力中小企业更好地利用人工智能代理 2026-07-21
Continued efforts are needed to make artificial intelligence infrastructure and tools more accessible to small and medium-sized enterprises, the nimble real-world adopters of the emerging technology, experts said at the 2026 World Artificial Intelligence Conference on Monday.
During a visit to the conference exhibition, Pu Yapeng, deputy director of the Shanghai Municipal Commission of Economy and Informatization, said he found many products developed by SMEs comparable to those by industry giants such as IBM and Huawei, thanks largely to affordable computing resources and government support.
"Computing power should become a utility, and tokens should be as affordable and easy to use as water or electricity. Only then can AI truly benefit more SMEs," he said.
Shanghai will continue to roll out compute, data and model vouchers for SMEs, expand engineer training to support the front-end deployment of AI agents and build dedicated AI innovation clusters for smaller businesses, Pu added.
SMEs are major drivers of AI innovation because of their flexibility and agility, said Li Sen, vice-president of Shanghai-based AI company Sense-Time. In China, innovative AI applications have already been widely adopted in e-commerce livestreaming, education, content creation, software development, marketing and back-office operations.
However, many SMEs struggle to identify the right AI tools amid the rapid pace of technological innovation, said Zhou Yun, a future industry development strategy expert at cloud and big data company Si-Tech Information Technology. Data security concerns, unpredictable infrastructure costs such as token expenses and steep learning curves for high-value AI solutions remain major barriers to wider adoption, Zhou said.
To address these challenges, the China Association of Small and Medium Enterprises launched a special initiative on Monday aimed at helping SMEs better leverage AI agents.
Industry associations such as CASME should take a greater role in curating, categorizing and standardizing AI solutions into ready-to-use "AI service packages", Li of SenseTime said. Such packages would enable SMEs to improve project performance more quickly while keeping trial-and-error costs low, he added.
Li Meng, president of the Chinese Society for Sustainable Development, said SMEs should avoid training foundational models and instead focus on vertical applications, seeking niche opportunities in fields such as industrial integration, pharmaceutical research and development, and energy.
However, because industrial and medical data often involve trade secrets and personal privacy, robust compliance frameworks must be established before such projects begin, he said.
Li also called on governments and financial institutions to adopt a long-term approach by introducing targeted action plans to support technology iteration and reduce AI adoption costs for SMEs.
According to a report released by market consultancy Accenture on Sunday, Chinese companies' digitalization index rose from 49 to 59 out of 100 this year, marking the biggest increase in nine years. However, only 10 percent of surveyed companies have successfully transformed their AI investments into sustainable long-term capabilities.
"The question companies face now is no longer whether to use AI, but how to turn it into real growth and lasting competitive advantage," said Samantha Zhu, chairperson of Accenture Greater China.
"That shift does not come from technology alone. It requires weaving AI deeply into organizational capabilities, core business processes and industry expertise. The data, institutional knowledge and handson experience that companies have accumulated over time will become the ultimate competitive advantage in the AI era," she said.
Source: China Daily
