The AI industry is in search of a new standard of measurement. As AI agents gradually enter real-world business scenarios, the focus is shifting to how many agents are genuinely being utilized and consistently creating value. Two months ago, Baidu founder Robin Li first introduced the DAA (Daily Active Agents) metric, aiming to gauge the development of AI applications by the number of agents that are truly active daily, complete tasks, and generate value. On July 17th, at the 2026 World Artificial Intelligence Conference (WAIC), IDC released the inaugural "DAA Research Report," further systematizing this metric.
The report indicates that the global number of active agents is projected to grow from 28.6 million in 2025 to 79.4 million in 2026, reaching 2.216 billion by 2030. This projection forms a key underlying theme for Baidu's showcase of multiple agent products at this WAIC. From Baidu Dazi for personal productivity to products like Baidu Wenku, Baidu Wangpan, and Miaoda, Baidu is actively seeking more high-frequency entry points for its agents.
This represents only one facet of DAA. For enterprises, the criteria for evaluating an agent's value are more direct: can it integrate into core business operations and deliver quantifiable efficiency gains or operational revenue? The Decision Agent Famou 2.0 exemplifies Baidu's exploration into how industries can apply AI.
Unlike agents for writing, summarization, and information organization in office settings, Famou targets complex decision-making scenarios such as enterprise production scheduling, process optimization, and logistics dispatching. It aims to assist companies in finding superior solutions under numerous constraints. Currently, Famou has been deployed across industries including ports, logistics, industrial manufacturing, chemicals, energy, and finance. It is also entering AI for Science domains like agricultural breeding and electromagnetic research.
According to Li Annan, product lead for Baidu Intelligent Cloud's Famou, the service currently caters to two main types of enterprises. The first category comprises large key account (KA) clients, including major manufacturers, energy companies, and port operators. Due to their large-scale operations, even a few percentage points of efficiency improvement can yield significant economic benefits. These clients typically opt for private deployment models, often involving project-based collaborations with a standard service cycle of about three years per project.
The second category includes industry niche leaders and regional frontrunners. These enterprises usually have annual output values ranging from 1 billion to 10 billion RMB and are better suited for public cloud subscription models. Compared to private deployment, the public cloud offers lower costs, thereby reducing the barrier for medium-sized enterprises to experiment with agents.
"Medium-sized enterprises are showing good acceptance now because public cloud pay-as-you-go models are much cheaper than private deployment, and they can see tangible results," Li Annan pointed out. Once an agent is integrated into an enterprise's production systems, its value must be measured through efficiency gains, cost reductions, and business revenue. Whether DAA can become the new standard of measurement for the AI era is now under close scrutiny.
Comments