Cloopen's Kong Miao on integrating AI into core business processes.

"In 2022, everyone regarded ChatGPT as just a product. In 2024, enterprise-level applications began to gain momentum, and manufacturers began to recognize some business value. This year, DeepSeek made large-scale models accessible for business-to-business use, and Manus also showed the new value proposition of Agents. All these have brought new inspirations." 

On July 28th, Cloopen AI hosted the "Reinventing Lift - AI Initiates New Growth" AI AGENT Practical Implementation and Industry Innovation Forum during the 2025 WAIC World Artificial Intelligence Conference. The forum focused on the technological evolution and industrial integration of AI Agent, and through vertical scenario solutions and global trend insights, it revealed how AI Agent can reshape the enterprise growth paradigm and promote the development of new quality productivity. 
Kong Miao, the vice president of Cloopen AI and the founder of Zhugu Intelligent Technology, gave a detailed presentation on the latest AI advancements and developments of Cloopen AI at the scene. 

He stated that Cloopen AI has always been driven by technological forward-thinking, always staying at the forefront of the industry in each iteration - from LLM Feature (embedding of large model functions) to Workflow (process/ orchestration-level applications) to Agent (autonomous planning and tool invocation), and truly integrating large models into the core processes of enterprises, making AI no longer just a tool but enhancing the efficiency and quality of the overall operational processes of the enterprises. 

"The delivery of results remains the core."

"This round of ChatGPT is different from previous ones. From computers to the previous generation of AI, they were all continuously developing productivity tools and waiting for people to use them. But this large model can make decisions on information and possesses some autonomous capabilities." 

Kong Miao stated in his on-site speech that from ChatGPT to Agent, technology is still evolving, including multi-Agent collaboration, other tool invocation ecosystems, etc. "But what remains unchanged is how to achieve the delivery of results, rather than providing a tool that is only useful when wanted and has no value when not used." 

The delivery of results not only relates to technological development, but is also closely linked to corporate decisions. Last year, Cloopen AI has released several specialized applications, beginning to integrate business knowledge and improving the efficiency of individual processes within these applications. 

This year, we also proposed earlier on how to assist customers in conducting reengineering, replacement, and designing new processes for the incremental market, helping enterprises achieve quality improvement and efficiency enhancement from a single point to business processes. On-site, Cloopen AI launched the Rongxi Agent & Copilot platform, providing enterprises with full-scenario empowerment covering marketing, customer service, quality inspection, and data insights through four intelligent engines: quality inspection agent, agent for customer service, agent for customer service, and insight agent. 

In her presentation, Kong Miao conducted case analyses for four typical scenarios. 

Quality inspection scenarios in the securities industry.
Previously, securities firms mainly relied on manual methods, which resulted in a long process, poor coverage rate and accuracy rate, and low ROI. For example, 5,000 calls might only have 1,250 valid ones. Finally, re-inspection by quality inspectors was required, and the overall coverage rate was also very low.
Last year, Cloopen AI released a large-scale quality inspection model. It set thresholds for some regulatory quality inspection requirements in advance. The large model could automatically mine through prompts or phrases, and could achieve analogical reasoning, significantly improving the operational efficiency and reducing the operational costs of enterprises. From the results, after adopting the large model, the quality inspection time could be shortened from 8 days to 3 hours, the coverage rate increased from less than 40% to 100%, the accuracy rate increased from less than 80% to 96%, and the missed detection rate dropped from 34% to 2%. 

Life insurance scenario.
The life insurance customer service team has a vast amount of customer communication data, but due to the limitations of previous technologies and resources, the utilization rate of the majority of this data is less than 5%. 

The solution provided by Cloopen AI is to integrate large models into the communication process for comprehensive data mining, simultaneously filtering out meaningful and structured information. Business personnel only need to ask questions based on the knowledge compiled by the large model. For example, analyzing the situation of financial management customers in the past month, which agents' soothing language was effective for customers, etc., and automatically inputting the data into the business system. This increases the data utilization rate to 95%, the scene coverage rate from the original 26% to over 90%, and shortens the analysis time from the original 10 days to 4.5 hours. 

Customer service after-sales scenario.
Take the bathroom industry as an example. Previously, after the customer contact center received a phone consultation, the agents had to consult very complex manuals, which was inefficient.
After adopting the ability of the Cloopen AI large model knowledge base, customers can build their own private domain large model knowledge base in advance, dig out good scripts and strategies. When customers raise questions, the agents can reply quickly. This change has shortened the after-sales call time from 12 minutes to 3 minutes, shortened the response time for complex problems from 30 minutes to 5 minutes, increased the problem resolution rate from 30% to 85%, and reduced the complaint rate from 35% to 7%. 

Banking scenario.
By adding more analytical tools and models, the marketing analysis cycle of banks has been shortened from the previous one month to one week or a few days.
In response to the lack of understanding of the analysis process by most enterprises, the new Agent can relocate typical industry scenarios internally. After receiving a problem, it first plans the tasks, automatically adjusts the analytical model, and provides conclusions and suggestions. This optimization upgrade has shortened the report analysis duration from 7 days to 3-4 minutes. 

Kong Miao stated that in the future, Cloopen AI will not only enhance business efficiency through large models, but will also deeply integrate its own communication infrastructure and data capabilities to build a "CC + CRM + AI + DATA" integrated resonance platform. This platform will connect the entire marketing, sales, and service chains of enterprises, bringing broader coverage, deeper scene penetration, and greater commercial value to enterprises. 

To provide enterprises engaged in maritime activities with a replicable intelligent operation model

Facing the long-standing problems in the financial industry such as insufficient professional capabilities, data silos and excessive manual work time, Wen Ge, the CTO of the Intelligent Product Department of Cloopen AI, presented a brand-new product - "Intelligent Business Analysis Handbook" at the scene. 

"This product is designed as an exclusive intelligent platform for the financial industry. It achieves automatic business insight through natural language interaction, reducing the strategy generation time from 7 days to 11 minutes and lowering the level of manual intervention to 9%," Wen Ge disclosed. Its dynamic knowledge graph integrates multi-source data such as the central bank's credit information and government affairs, and builds a closed-loop decision-making capability for autonomous correction and execution, precisely addressing the "last mile" pain point for city commercial banks from data to growth. 

Relevant data shows that the global AI Agent market is expanding at a compound annual growth rate of 44.8% to reach a scale of 47.1 billion US dollars by 2030. Therefore, Cloopen AI will continue to drive the technology to evolve towards "environmentally adaptive intelligent agents", deepen vertical application scenarios, and reshape industry boundaries with new quality productivity. 

Simon Amran, the Director of Digital Intelligence Solutions for Southeast Asia, shared at the forum that the region, with its 700 million young population, 80% smartphone penetration rate, and a projected $1 trillion digital economy by 2030, has created a huge demand for AI. It is expected that by 2025, 70% of Southeast Asian enterprises will deploy AI solutions. 

In response to the challenges such as high costs of multilingual services and a sharp increase in consultation during promotional periods, Cloopen AI Virtual Agent achieves 80% of consultations to be automated through dialect understanding, cross-channel integration and flexible architecture. The response time is reduced from 3 days to 10 seconds, providing a replicable intelligent operation model for overseas enterprises. 

On the spot, the cooperation between China and South Korea showcased the advantages of South Korea's hardware research and algorithm optimization, as well as the new path of synergy and complementarity with China's market resources and the ability to implement scenarios. The "Zhongguancun Intelligent Digital Artificial Intelligence Industry Alliance International Cooperation Working Committee" that was simultaneously inaugurated operates in a tripartite model of cross-border technology collaboration, policy research, and project incubation, helping Chinese enterprises break through barriers in overseas markets.

"The most profound challenge is not the technical bottleneck, but how humans can coexist with intelligent entities that are more 'efficient' than themselves." Gu Ha, the vice president and secretary-general of the Zhongguancun Intelligent Digital Artificial Intelligence Industry Alliance, emphasized that the rise of AI Agents "starts with technology, is achieved through industry, and culminates in civilization."