AI Agents Will Serve People, Not Devices, Says Lenovo Group CEO

AsianFin -- Lenovo launches its full-scenario "Super Intelligent Agent Matrix" at the company's annual Tech World conference on Thursday.

The new lineup includes intelligent systems for individuals, enterprises, and cities, as well as an advanced AI inference engine developed in collaboration with top research institutions.

Lenovo Group Chairman and CEO Yang Yuanqing declared the era of "device-bound agents" over, stating that super intelligent agents will instead "bind with people, not hardware," making every individual the owner of a dedicated AI system.

Eight years into its AI journey, Lenovo is pivoting from traditional hardware to cognitive infrastructure. Building on last year's launch of the Xiaotian AI PC assistant, Lenovo has now introduced a new class of AI: the Super Intelligent Agent, defined by three core capabilities—perception and interaction, cognition and decision-making, and autonomy with evolution.

"This is no longer just a tool," said Yang. "It's a cognitive operating system for individuals and enterprises."

The personal version, known as Tianxi, enables seamless access across devices to calendars, emails, and personal cloud storage. It autonomously orchestrates tasks based on real-time understanding of user needs. The system, already available on four Lenovo AI smart terminals, will be upgradeable on existing devices that meet hardware standards.

Sales of Lenovo's AI PCs have already crossed the one-million mark, with performance metrics such as user satisfaction, engagement, and Net Promoter Scores showing steady improvement, according to Liu Jun, Executive VP of Lenovo Group and President of Lenovo China.

Lenovo also launched Lexiang, its enterprise-focused super intelligent agent, designed to integrate across sales, procurement, and supply chains. Built on a four-layer "device-edge-cloud-network" architecture, Lexiang taps into hybrid cloud infrastructure and leverages private data to create vertical AI solutions.

"We aim to replace 30% of our workflows with silicon-based employees this year," Liu said.

To support this transformation, Lenovo has built its own AI application library and model factory, and plans to expand intelligent agents across industries by applying general-purpose and vertical AI models.

Beyond individuals and enterprises, Lenovo is targeting cities with its "1×N" intelligent system architecture. In pilots across Wuyishan and Yichang, the Lenovo city agent has demonstrated early success in digital governance and social services, positioning the company as a key player in China's AI-powered public infrastructure.

With the rapid expansion of edge AI, Lenovo emphasized the need for both lightweight models and robust local processing power. The company announced the Lenovo Inference Acceleration Engine, co-developed with Tsinghua University and Wuxin Chip Research, to drive hardware-software synergy.

This engine boosts performance for edge-side AI with faster preloading, lower memory usage, reduced energy consumption, and shorter inference times—making AI more efficient and sustainable.

Yang estimates that edge AI capabilities will triple in the next 12 months, thanks to advances in both chips and models.

As super intelligent agents access more sensitive personal and corporate data, Lenovo stressed that security and privacy are "non-negotiable values." Its upcoming AI platforms will integrate deepfake detection and employ trusted computing frameworks across personal and cloud environments.

Addressing growing fears around AI replacing human jobs, Yang offered reassurance: "There's no need to worry about being replaced by AI—but standing still is not an option."

Instead, he called on individuals and organizations to master AI tools and "continuously break self-imposed limits." He noted, "In the AI era, those who can leverage technology best will be the ultimate winners."