Only by accelerating response can faster growth be achieved.
In the short drama market, the golden window for a hit is extremely short. Whether a series can capture traffic and drive paid conversions largely depends on whether the front-end product has undergone repeated fine-tuning.
For Shanhai Xingchen, the mini-program is the central scene where change occurs.
Founded less than three years ago, Shanhai Xingchen has already risen to become one of China's leading short drama teams, continuously operating over 500 short drama IPs and their corresponding mini-programs. Mini-programs connect content distribution, user conversion, and monetization—every optimization in activity configuration or payment experience could influence a drama’s market performance.
High-Frequency Iteration Challenges Collaboration Efficiency
High-frequency iteration tests collaboration efficiency among content, operations, product, R&D, and testing teams. Feishu Projects AI and Feishu CodeM are precisely the key tools supporting this collaborative chain.
The Faster Business Accelerates, the Sooner Legacy Tools Hit Bottlenecks
In the short drama race, speed is the foundation of survival. The lifecycle of a hit is often measured in weeks or even days. From script review and production planning to mini-program development and testing acceptance, every phase competes against time. Any delay means real financial loss.
In its early days, Shanhai Xingchen relied on manual efforts to sustain growth. When there were only a few projects, creating chat groups, holding meetings, and circulating documents was manageable. However, as concurrent projects increased from a handful to dozens, the human-driven "chasing" model quickly approached its limits.
Ouou, the testing lead, was the first to feel the pressure. As part of the downstream delivery team, she was highly sensitive to information gaps. Behind seemingly normal processes lay massive hidden inefficiencies. For example, writing a PRD typically took two hours; for complex requirements, repeatedly switching between Feishu Projects and third-party tools to organize and re-enter data not only increased the risk of data leakage but also consumed significant time.
A more common bottleneck occurred after requirements entered the R&D phase: while product teams synced requirements, developers had to switch back locally to check code, pull branches, and confirm modification scopes—rebuilding context before development could begin. The more projects there were, the longer the waiting time.
The team realized that their existing toolset could no longer meet the needs of real-time decision-making.
Feishu CodeM, Embedded into the R&D Process
What truly allowed Shanhai Xingchen to experience seamless R&D integration was the deep incorporation of Feishu CodeM.
As a native Feishu AI-powered R&D agent, CodeM functions as a coding agent tailored for enterprise R&D scenarios, seamlessly integrating into Feishu and Feishu Projects’ collaboration workflows to handle tasks such as feature development and bug fixing.
When a requirement flow reaches the "AI Development" stage, Feishu CodeM automatically reads the requirement and technical plan, performs requirement analysis, code modifications, and compilation verification within a cloud sandbox, then automatically submits a merge request (MR), synchronizing updates to Feishu Projects and advancing the task to the next stage.
The entire process requires no manual intervention and does not occupy local environments. Code runs in an isolated sandbox, allowing developers to focus solely on MR reviews—concentrating their energy on strategic judgment rather than repetitive environment setup and context rebuilding.
This means frontend changes related to activity configurations, payment optimizations, or new drama launches can continuously progress through the workflow without stalling at manual handoffs.
For short drama operations, this process not only saves coding time but, more importantly, reduces context-switching delays, tool transitions, and manual transmission lags during development. As AI continuously works along the workflow, the team’s core focus shifts from execution to judgment and decision-making.
Making AI an Integral Part of the Workflow
The transformation of the R&D agent from “waiting for instructions” to “scheduled within the workflow” stems from Shanhai Xingchen’s AI-driven workflow system built on Feishu Projects.
Wherever requirements flow, AI participates—engaging throughout stages including requirement drafting, technical review, testing validation, defect analysis, and code development. Tasks that once required manual initiation and advancement now run automatically within the workflow, enabling full AI integration across the entire product-R&D collaboration chain.
After a mini-program requirement is created, Feishu Projects automatically triggers an AI node that generates a structured PRD cloud document based on background, objectives, and constraints. Requirement approval and document generation are nearly completed simultaneously, reducing PRD drafting time from two hours to five to ten minutes, with 100% completeness.
During technical reviews, AI proactively checks the completeness of plans, identifies missing elements, and suggests revisions. Review meetings no longer get bogged down in “what wasn’t clearly written,” but instead focus on critical design discussions, saving one to two hours per session on average and enabling faster entry into development for new drama launches and campaign adjustments.
In the testing phase, AI automatically executes test cases and generates structured acceptance checklists, identifying issues requiring manual confirmation in advance. Even when defects arise, AI produces reports containing problem localization, root cause analysis, verification steps, and resolution suggestions. In one iteration, 162 defects were processed using standardized procedures, reducing manual handling time by approximately 50%.
Ouou believes the key lies in Feishu Projects’ low barrier to native AI configuration—AI orchestration is “as simple as configuring a field.” This enables teams to quickly customize workflows according to business scenarios, embedding large model capabilities directly into workflow nodes while avoiding data leakage and redundant information transfer.
Smoother Workflows Lead to Easier Management
The impact of AI extends beyond coding and defect analysis. As it becomes embedded throughout the entire process, many management tasks previously requiring repeated confirmation, aggregation, and judgment now run automatically.
Every morning, managers receive an automated project insights report generated by the Feishu Projects assistant “Mili.” The AI summarizes progress across more than 20 ongoing projects, identifies resource bottlenecks and review congestion, helping managers quickly pinpoint issues without manually checking each item.
Workforce analysis has shifted from “post-hoc statistics” to “preemptive planning.” Project managers can initiate AI-powered analysis via natural language queries to rapidly understand team workload, resource distribution, and potential conflicts—identifying risks before requirements kick off and enabling more reasonable scheduling.
AI has also improved data quality. Take the change reason field as an example: initially, only options A, B, and “Other” existed. After adding C and D, historical records remained piled under “Other.” Now, AI can reclassify ambiguous data based on business semantics, restoring value to legacy data.
This continuous, online data accumulation has also transformed project retrospectives. The system automatically aggregates data on scheduling, working hours, defects, and documentation, generating retrospective reports that include execution summaries, progress analysis, and improvement recommendations.
At Shanhai Xingchen, AI is no longer just a tool for completing tasks—it has become an organic component that continuously participates in workflows, allows data to accumulate naturally within processes, and enables management actions to emerge seamlessly from workflows.
Accelerating Response Enables Faster Growth
Time saved at each workflow stage ultimately translates into business advantage. With AI integrated into workflows, content teams can validate market feedback faster, operations teams adjust campaigns and pages promptly, R&D responds swiftly to requests, and testing identifies risks earlier—significantly enhancing overall responsiveness.
Zhou Peijin, Chairman of the Board and CEO of Shanhai Xingchen, said: “Feishu Projects has injected digital momentum into short drama production. From content creation to quality testing and cross-functional collaboration, end-to-end digitization has enabled us to achieve our strategic goals of ‘improving efficiency, accelerating speed, ensuring quality, and reducing costs.’ Enhanced labor efficiency and zero-loss quality assurance are key to maintaining our market leadership.”
These transformations have yielded measurable results: through Feishu Projects, Shanhai Xingchen has achieved full digitalization from content production to post-production, reduced project document retrieval time to as fast as one minute, cut cross-departmental communication time by 50%, improved demand delivery efficiency by over 30%, and saved the equivalent of approximately 19.2 person-years in communication and coordination costs annually.
In the short drama industry, time itself is competitiveness. When workflows operate smoothly, teams can dedicate more energy to the creative, content-related, and strategic decisions that determine a drama’s success. The value of Feishu Projects AI and Feishu CodeM lies precisely in pushing every requirement flow, collaboration, and delivery one step further.
If you're also interested in use cases for Feishu CodeM, click the link in the lower right corner to get your exclusive solution!