"AutoAgents", which wants to be an enterprise-level intelligent body, received tens of millions of yuan in angel round financing | New projects emerge

Written by

Silas Grey

Updated on:July-13th-2025

In the future, expert agents may reach a “winner takes all” level.

One sentence introduction

Based on the self-developed Multi-Agent architecture, we provide enterprises with intelligent service providers that can be deployed in the production process.

About the Team

Yang Jinsong (CEO): Former Director of Product/Commercialization at DAMO Academy, former head of AI at ByteDance Feishu and head of Amazon AWS aPaaS platform, led the launch of Ali Lingjie and Tongyi-Alicemind, and managed product revenue of over 2 billion. Dr. Wang (Chief Scientist): Ph.D. from Columbia University, former research scientist at Alibaba DAMO Academy and Google Research, cited 12,000 times by Google Scholar.

Other core team members come from Alibaba DAMO Academy, Tencent, ByteDance, Amazon AWS and Google, etc.

Financing progress: Recently completed tens of millions of yuan of angel round financing, led by Linke Venture Capital, followed by Jimen Asset Management, and old shareholder Sinovation Ventures continued to follow up. This round of financing will be mainly used for product research and development and market expansion.

Products and commercialization

In the domestic market, AutoAgents, through its "Lingda" platform Agent Builder, targets the needs of the enterprise-side Agent market and provides mature technical solutions to solve issues such as data security, permission management, and system integration that enterprises are concerned about when applying large models.

At present, AutoAgents products have served leading customers in the power, finance, Internet, manufacturing and other industries, building a human-AI hybrid workflow for enterprises, greatly improving the operational efficiency and innovation vitality of enterprises. The company obtained tens of millions of commercial contracts in 2024 and is currently the Agent application product with the largest market share in the power industry, serving the State Grid and more than ten of its subsidiaries.

The company's products have been introduced by 5 cloud vendors, providing more than 100 open and closed source, field-vertical models, and promoting products on a large scale through more than 20 industry partners.

In overseas markets, AutoAgents has also launched a standardized product, Agents Pro, a social media operation tool for SMBs that adopts a free trial and community communication model.

Unlike Coze, Dify and other To C Agent platforms, AutoAgents differentiates itself by focusing on the enterprise market. AutoAgents does not simply provide tools, but helps enterprises deliver service results and achieve "paying for results" through standardized Agent products and industry solutions, such as providing more fine-grained permission management, data dashboards and database docking capabilities, and hybrid cloud/all-in-one machine deployment solutions.

To achieve enterprise-level availability, AutoAgents integrates enterprise-level RAG, AI Coding, Text2Agent, and visual workflow components. In terms of deployment, AutoAgents supports hybrid cloud/all-in-one machine deployment and can adapt to domestic computing power.

At the bottom of the product, Lingda introduces a unique multi-agent collaboration mechanism that can solve the context window limitation in the multi-step reasoning process; it supports generating Agent applications from a sentence, which can greatly improve the efficiency of development and deployment. During the task process, Lingda can decompose complex tasks and assign them to different professional agents, and the coordinating agent will coordinate the work of each part.

Currently, Lingda can also help humans "do real work". AutoAgents supports Anthropic's MCP protocol, allowing intelligent agents to discover and call external tools more efficiently.

Lingda can also simulate human-operated computing (similar to OpenAI Operator). With the help of the built-in Docker sandbox system, the intelligent agent can independently browse the web, retrieve data, and call commonly used software to complete designated tasks.

After the DeepSeek craze this year, the market generally believes that this is a critical moment for the "first year of Agents". AutoAgents has also developed a "Meta-Zhi" assistant for developers and individuals, which will be launched in 2025. This is an autonomous intelligent product that can run in real business scenarios and complete research and analysis tasks in professional fields on its own.

With agent fine-tuning technology, AutoAgents has been able to enhance the tool calling capabilities of intelligent agents, optimize collaboration efficiency, and improve code generation quality. AutoAgents has accumulated more than 20 patents, software copyrights and other intellectual property achievements in this field, and has published papers at top international conferences many times.

After this round of financing, AutoAgents will continue to rapidly commercialize and launch products for the To C market, and also plans to expand into overseas markets.

Founder’s thoughts

The core difference between enterprise-level Agent and personal Agent is that the former needs to meet the stringent requirements of enterprises in terms of data security isolation, authority system stratification, and deep system integration. General Agent focuses on ease of use and versatility, while enterprise-level Agent requires deep customization to adapt to complex business scenarios.
The current implementation of Agent technology still faces many challenges. Even if inference models such as DeepSeek R1 have powerful capabilities, in actual enterprise applications, a lot of engineering transformation is still required, deep adaptation with the existing tool chain, and integration of small models in the field, in order to effectively control hallucinations and ensure the reliability and security of output results.
The final outcome of Agent may be a "winner takes all" situation. In a specific vertical field, intelligent agents that can effectively accumulate industry know-how, precipitate best practices, and use high-quality data for in-depth training will build a deeper competitive barrier.
The future development direction of Agent is to re-plan enterprise workflows through deep collaboration between AI experts and industry experts, automate complex workflows, and enable humans to focus on high-density decision-making and responsibility. AutoAgents will develop into a "one plus N" business model in the future, that is, to produce various Agent products and solutions through a technology platform, realize pricing based on service volume, and obtain sustainable benefits from the global service value chain, which can break through the revenue ceiling of traditional software sales.
Enterprise software is shifting from "paying for tools" to "paying for results". The core value of Agent lies in service results orientation. Through standardized Agent products and industry solutions, it directly creates business value rather than simply providing tools. The fundamental purpose of enterprises choosing Agent is to solve actual business problems, rather than just paying for models.
Reasoning models represented by DeepSeek R1 have significant value in broadening the ability of agents to solve open-ended problems, and are particularly suitable for C-end applications such as code writing and novel creation. However, in enterprise-level applications, it is still necessary to carefully evaluate its scenario applicability, security alignment capabilities, and illusion control levels to avoid having overly high expectations of model capabilities.

What does "Intelligence Emergence" mean?

Agents are already a topic that cannot be ignored in 2025, but the track is still in its early stages. The core competitiveness of AutoAgents lies in their ability to achieve availability in enterprise-level production environments. They have been successfully implemented in large enterprise scenarios with low fault tolerance, such as power finance.

AutoAgents has a clear understanding of the current large model capabilities and boundaries. In different industries, it has both enterprise-level Agents products and SaaS products for emerging business scenarios such as social marketing in overseas markets. The company not only entered the market early, but also quickly sought commercial implementation. Both the product matrix and market strategy are relatively clear.