本次调研面向LGECH全体部门发放问卷，共回收125份有效回答。 This survey was conducted across all LGECH departments and collected 125 valid responses.
问卷共10题，按"任务强度评估(Q1) → AI效率提升需求(Q2-Q6) → 困难与改进建议(Q7-Q10)"三层递进设计。 The survey consists of 10 questions in three stages: task intensity assessment (Q1) → AI efficiency needs (Q2-Q6) → challenges and improvement suggestions (Q7-Q10).
| 1 | 数据处理与分析（数据导出、数据核对、报表制作、数据分析等）Data Processing & Analysis (data export, data validation, report creation, data analysis, etc.) |
| 2 | 文档/PPT/内容创作（写邮件、PPT制作、图文视频制作等）Document/PPT/content creation (email writing, PPT creation, image/video production, etc.) |
| 3 | 市场信息调研（竞品信息、价格跟踪、社媒舆情、信息采集等）Market Information Research (competitor information, price tracking, social media monitoring, information collection, etc.) |
| 4 | 流程与系统操作（打禀议，权限配置，提系统流程，审核凭证等）Process & system operations (approval requests, permission setup, workflow submission, voucher review, etc.) |
| 5 | 跨语言沟通与翻译（外语邮件翻译，跨语言会议，口语交流等）Cross-language communication & translation (foreign-language email translation, multilingual meetings, verbal communication, etc.) |
| 6 | 其他业务场景 Other business scenarios |
Q1结果显示数据处理业务占整体工作时间的40%以上，77名受访者表示在该业务上花费时间最多。 Q1 shows that data processing accounts for over 40% of overall work time, and 77 respondents said they spend the most time on this work.
导入AI后预计可平均缩短95%以上的时间，需要实现数据提取→清洗→分析→报告生成全流程自动化的解决方案。 With AI adoption, average time savings of over 95% are expected, requiring an end-to-end solution that automates data extraction → cleansing → analysis → report generation.
68%的受访者每天在多语言沟通上花费1h以上，特别是对实时会议口译和多格式文档翻译的需求很高。 68% of respondents spend more than one hour per day on multilingual communication, with especially high demand for real-time meeting interpretation and multi-format document translation.
需要摆脱现有的单纯文本翻译工具，构建支持视频/语音/图片等多种媒介的综合翻译平台。 A comprehensive translation platform supporting video, audio, images, and other media is needed beyond simple text translation tools.
Q9结果显示员工最偏好的AI工具是使用简单、可及性好的低门槛工具，多数意见认为"比起复杂功能，更需要简单实用的功能"。 Q9 shows that employees prefer easy-to-use, accessible AI tools, with many comments emphasizing simple and practical functions over complex features.
导入AI时，"先让员工用起来"最重要，需要采取分阶段扩展功能的思路。 When introducing AI, the most important step is to get employees to start using it, then expand functionality gradually.
| 优先级 Priority | 行动建议 Action Recommendation | 依据 Basis |
|---|---|---|
| 1 | 优先解决网络访问与业务适配双重瓶颈 Resolve the dual bottlenecks of network access and business adaptation | Q8前两大困难为"业务场景复杂，AI难以适配"（80人）和"无法连接境外服务"（69人）。需同时推进内网AI部署与业务场景定制化。 In Q8, the top two difficulties were "business scenarios are complex and AI is hard to adapt" (80 people) and "unable to access overseas services" (69 people). Internal AI deployment and business scenario customization should be advanced together. |
| 2 | 数据处理全链条自动化 Automate the full data processing chain | Q1任务强度最高（77人选"多"），Q2需求覆盖"导出→清洗→核对→报表→洞察"全链路，单点解决方案不足。 Q1 shows the highest task intensity for data processing (77 selected 'High'), and Q2 needs cover the full chain of export → cleansing → validation → reporting → insights, so single-point solutions are not enough. |
| 3 | 升级翻译工具至多媒体层面 Upgrade translation tools to a multimedia level | Q6翻译需求已从文本扩展到"多格式内容"（73人次）和"跨语言会议"（67人次），简单文本翻译工具已不够。 Q6 translation needs have expanded from text to "multi-format content" (73 people) and "cross-language meetings" (67 people), so simple text translation tools are no longer sufficient. |
| 4 | 以"低门槛"为第一原则推广 Promote low-barrier use as the first principle | Q9中豆包、DeepSeek等低门槛工具提及最多，Q10反复出现"简单实用"——先让员工用起来，形成习惯后再深化。 In Q9, low-barrier tools such as Doubao and DeepSeek were mentioned most often, and Q10 repeatedly emphasized "simple and practical." First encourage employees to use AI, then deepen features after habits are formed. |