Zhibo8: How an AI programmer boosts R&D efficiency

Zhibo8 leverages Qoder CN to enhance its AI-assisted collaborative development, overcome efficiency bottlenecks, and improve code quality. Since its founding in 2007, Zhibo8 has been a leader in sports event broadcasting and information services. As a well-known sports broadcasting and information platform in China, Zhibo8's mission is to make it easier for sports fans to obtain event information and watch live streams. Through its app, website, and other channels, Zhibo8 provides comprehensive services such as live sports broadcasts, news, and data analytics. These services cover a wide range of sports, such as football, basketball, tennis, and esports, bringing the digital sports experience to users. Zhibo8 has built strong user loyalty by attracting many sports fans to communicate and discuss events on its platform. The platform has over 250 million users, with its app exceeding 50 million monthly active users and 13 million daily active users. Its user loyalty and popularity have led the industry for many years. Zhibo8 also continuously improves its platform performance and user experience through IT investment and strategy. This includes research and development (R&D) and innovation, a mobile-first focus, system architecture optimization, data security and privacy protection, content integration and copyright management, and community interaction. These efforts help solidify its leading position in the sports broadcasting and information services industry.

​

Choosing tools in the era of large models

As a top sports broadcasting platform in China, Zhibo8's development team supports frequent business iterations and technical optimizations for high-concurrency scenarios. The team had previously used ChatGPT for technical Q&A and code snippet generation. They also allowed team members to choose their own AI coding assistants. However, this approach led to several pain points:

1. Low development assistance efficiency

When using GPT, the lack of project context meant developers had to ask continuous follow-up questions to improve code accuracy. This process was time-consuming and reduced development efficiency. Additionally, most AI assistants on the market could not support the various programming languages and preferred IDEs used by the team's developers. This forced team members to switch between tools and IDEs, wasting valuable time.

2. Poor business adaptation

Zhibo8 has well-established coding standards and architectural designs. The code generated by open source and free AI programming assistants was often incompatible with the company's standards and business scenarios, requiring frequent manual adjustments.

3. Lack of efficiency metrics

Before standardizing on a single AI programming assistant, team members used a fragmented set of tools. This fragmentation made it impossible to centrally manage AI tool usage or measure the impact on the team's overall efficiency. It also hindered collaboration.

​

Address your pain points, choose Qoder CN

Zhibo8 chose to adopt Qoder CN Enterprise Dedicated Edition to address its pain points related to R&D efficiency, code quality, knowledge management, and data security.

1. Seamless integration for an upgraded developer experience

Qoder CN supports mainstream IDEs, such as VS Code, JetBrains IDEs, and the Visual Studio series. Developers do not need to switch their development environment or install additional plugins, which significantly reduces learning and migration costs. Teams can quickly use out-of-the-box features, such as smart code completion, code explanation, unit test generation, and the AI programmer.

2. Adherence to enterprise-level code standards

By uploading the enterprise codebase to the Qoder CN knowledge base, the AI-generated code automatically adheres to Zhibo8's coding standards and system architecture, which reduces the manual review workload. For example, when developing a live streaming throttling system, the code generated by Qoder CN directly matched the exception handling mechanism established by the team, resulting in an overall code adoption rate of over 20%.

3. End-to-end AI assistance for complex scenarios

Qoder CN not only supports line-level and function-level code completion but also leverages its AI programmer to understand and manage multi-file project contexts. This assists developers with complex tasks, such as implementing requirements, fixing bugs, and performing local code reviews. When refactoring their live comment service, the Zhibo8 team used this feature to quickly identify legacy code logic and generate optimization plans, which reduced their development cycle by 50%.

​

Actual results

Zhibo8 achieved significant improvements in both efficiency and quality after adopting Qoder CN.

• Boosted development efficiency: The Qoder CN AI coding assistant provides features such as code completion, multi-file code modification, and autonomous task execution. These features helped the team save approximately 30% of their coding time on standard feature development and increased the efficiency of unit test generation for complex modules, such as real-time data synchronization, by 30%.
• Improved code quality: Code that adhered to company standards increased the first-pass rate by 35%. Comment coverage rose from 30% to over 70%.
• Faster knowledge onboarding: With the code explanation and flowchart generation features of Qoder CN, new team members could quickly understand the legacy system architecture, which shortened their training period by 30%.

Qoder CN Enterprise Dedicated Edition helped Zhibo8 implement R&D standards, improve team efficiency, and protect data assets by creating a closed-loop process that incorporates knowledge fusion, intelligent generation, security controls, and efficiency quantization. For enterprises pursuing intelligent transformation, this represents not just a tool upgrade, but a strategic leap toward a human-AI collaborative R&D model.