Baidu's Shift: Focusing on AI Agents to Transform Productivity and Interaction
At the Baidu AI Developer Conference, Create 2026, held in Beijing, the focus sharply turned towards intelligent agents as the centerpiece of the AI evolution. During the keynote, Robin Li, Baidu's founder, articulated a significant transition from competing on the scale of AI models to competing based on the capabilities of task-executing agents.
Li emphasized that the surge of interest in AI products has pivoted this year. “The real catalyst for AI’s viral growth was not the foundational models but the innovative applications built atop them,” he noted. This shift indicates that users are less concerned with whether AI can "think" and more focused on its ability to accomplish tasks effectively.
The core of Li's vision is the evolution of AI from simple chatbots into sophisticated digital workers capable of taking on complex workflows. These agents are not merely designed to respond to commands but are built to operate independently, continuously learning from their environments while iteratively refining their processes.
Self-Evolution Across Levels
Li introduced the notion of self-evolution as a transformative aspect of AI, which unfolds across three domains: intelligent agents, individuals, and enterprises. For agents, this means advancing beyond passive roles; future AI will possess the capacity for error correction, learning from mistakes, and optimizing its responses without human oversight.
As this technology grows smarter, the traditional dependency on human input will diminish, paving the way for a new model of interaction and productivity.
Emergence of the Super Individual
Turning to the individual level, Li posited that AI ushers in a new era of “super individuals,” where the smallest unit of productivity within organizations morphs from teams to individuals augmented by fleets of intelligent agents. With advances in code generation and task execution, everyday creators and developers will experience a marked enhancement in their productivity potential.
Transforming Enterprise Management
At the organizational level, AI is set to flatten corporate hierarchies. Li predicted a transition in management styles, where oversight evolves from direct supervision to more collaborative approaches centered around alignment and empowerment instead of control.
However, he acknowledged that the greatest hurdle organizations will face isn't just implementing AI but adeptly integrating data and workflows to maintain a verifiable, continuously optimizing framework.
Introducing Daily Active Agents
Baidu’s concept of Daily Active Agents (DAA) takes inspiration from Daily Active Users (DAU) of mobile platforms. Li argued that the value should not be measured simply in terms of engagement costs but by the number of agents effectively completing tasks. Looking ahead, he projected that the global tally of active agents could surpass 10 billion.
The evaluation of platform success will hinge less on user engagement time and more on the productivity output facilitated by these agents.
Impact on Software Development
Li further explored the repercussions of AI advancements on the software sector. With enhanced code generation capabilities, the barriers to software development are crumbling, fostering a market geared towards disposable applications. He suggested that soon, users might customize and generate specific software on the fly, further pivoting the software industry towards an on-demand model.
New AI Products Unveiled
At the conference, Baidu showcased a suite of new AI products designed around the agent theme. The launch of DuMate, a versatile AI agent, highlighted its capabilities in various areas: from resolving customer service issues to data analytics and content generation.
The Miaoda app, including its enterprise edition, emerged as a prominent code-generation agent, reportedly able to autonomously produce approximately 90% of its code. Moreover, Baidu introduced Baidu YiJing, a multi-agent platform tailored for digital humans, enhancing capabilities in live streaming and interactive media.
Baidu’s shift illustrates a foundational change in AI’s trajectory—one that could reshape productivity and strengthen human-agent collaboration in both workplaces and creative fields.
