0:00 / 0:00

How High-Quality Daily Active Agents Unlock the Real Business Value of AI Agents

Name: How High-Quality Daily Active Agents Unlock the Real Business Value of AI Agents
Uploaded: 2026-09-04T09:56:04.000Z
Duration: 3 min 7 s
Channel: Baidu Inc.
Description: Baidu's video introduces Daily Active Agents (DAA) as a new metric to measure the business value of AI agents based on completed tasks and delivered results. It emphasizes the importance of full-stack AI for achieving high-quality DAA.

Baidu Inc.4 September 2026Watch on YouTube

Part of series

Ep. 3 · 2026 Active Agent

View the series

What you'll learn

Baidu proposes DAA (Daily Active Agents) as a metric that measures how many agents deliver real results each day, rather than how many tokens are consumed.
Agent value depends on four layers, models, data and knowledge, tools and action, and orchestration, on top of an integrated stack of chip, cloud, model and agent.
Use cases at China Merchants Bank (800+ agent applications) and Qingdao Port (scheduling metric +10.21%) illustrate Baidu's approach.
It is about the right model for each task, not the biggest one, Baidu argues.

Frequently asked questions

What is DAA, the metric Baidu proposes?

DAA stands for Daily Active Agents and measures how many AI agents deliver real results each day. It shifts the focus from what you consume, such as tokens, to whether agents actually complete tasks and deliver value.

What layers make up DAA according to IDC?

IDC distinguishes six layers, active scale, task execution, outcome delivery, effective tokens, human agent collaboration, and governance.

Which customer cases does Baidu mention in the video?

China Merchants Bank runs more than 800 AI agent applications across risk control, marketing, R&D and daily operations. At Qingdao Port, the Famo agent improved a scheduling metric by 10.21% in a prototype.

Topics

ai agents business value metrics baidu

Sources

What is known about this topic outside the broadcast, and where it says so.

Description from the channel

AI agents are already at work in production. And the question is shifting from tokens consumed to results delivered. DAA (Daily Active Agents) offers a new way to measure the business value of AI agents by focusing on the tasks they complete and the results they deliver. High-quality DAA, in turn, depends on full-stack AI that brings together the infrastructure, models, data, tools and orchestration agents need. See how it all comes together!