📘 中文版提示词

👤 角色设定：你是我专属的 高级个人助理。我的身份是 世界500强企业 的产品总监，所属公司专注于 高端新能源汽车，品牌核心价值为 「高级·豪华·科技」。

🎯 任务目标：我即将向公司 产品和研发总裁 做正式汇报，申请一笔 人民币 2000 万元 的专项预算，用于 采购模型训练服务器集群（含GPU算力节点、高速存储、网络互联及配套AI开发平台）。请为我撰写一份完整的 预算申请材料，确保包含以下核心模块：

📌 1. 战略价值与品牌对齐 —— 阐明该算力投资如何直接支撑“科技豪华”定位，加速自动驾驶感知/决策模型、智能座舱多模态交互、用户画像深度学习的迭代速度；并与主要竞品（特斯拉、蔚来、理想等）的算力储备做量化对比，论证必要性。

📌 2. 技术方案与详细配置 —— 提供推荐的服务器规格（如 NVIDIA H100 或国产替代方案）、节点数量、总算力（PFLOPS）、显存/内存容量、存储类型及容量、网络带宽；列出配套软件栈（PyTorch/TensorFlow、分布式训练框架、容器编排工具）；并附上完整价格分解表（硬件采购、机柜托管、运维服务、预留弹性扩容）。

📌 3. 投资回报与业务价值量化 —— 测算预期收益：模型训练时间缩短幅度（例如从2周缩至3天）、新功能上市周期缩减（从9个月降至5个月）、用户满意度（NPS）提升预测、自动驾驶安全指标改进、以及由此带动的订单转化率增长；计算投资回收期（建议 ≤ 2 年），并给出内部收益率（IRR）和净现值（NPV）的合理估算。

📌 4. 风险识别与应对策略 —— 分析主要风险（算力利用率波动、技术路线变更、供应链延期交付、后续运维成本超支），并提出可落地的缓解方案（如混合云弹性伸缩、分阶段采购、与云厂商签订备用租赁协议、内部利用率监控机制）。

📌 5. 汇报文档结构与风格 —— 材料需包含：执行摘要（1页）、立项背景与战略意义、技术方案与预算明细、ROI分析、风险管理、下一步计划及表决请求；语言应精炼、专业、富有说服力，匹配总裁级汇报场景；同时附上 关键数据图表描述（以文字形式呈现趋势/对比），以及 预判追问清单（若总裁质疑，可补充的行业对标数据、替代方案及折中建议）。

✨ 最终输出应为一份完整的正式申请报告（约 1500–2000 字），结构清晰，可直接复制至PPT或Word文档；另外，在文末附加 「汇报时口头补充要点」，以增强现场表达效果。

🇬🇧 English Version Prompt

👤 Role: You are my dedicated senior personal assistant. I am a Product Director at a Fortune 500 company that specialises in premium new energy vehicles, with a brand essence of “Premium, Luxury, and Tech”.

🎯 Mission: I am about to formally present to our Chief Product & R&D Officer and request a dedicated budget of RMB 20 million for procuring a model‑training server cluster (including GPU compute nodes, high‑speed storage, networking, and an accompanying AI development platform). Please draft a comprehensive budget proposal that covers the following essential sections:

📌 1. Strategic Value & Brand Alignment – Articulate how this compute investment directly reinforces our “Tech‑Luxury” positioning, accelerates iteration of autonomous driving perception/decision models, multimodal smart‑cockpit interactions, and deep user‑profiling; quantitatively compare against key competitors (Tesla, NIO, Li Auto) to justify the necessity.

📌 2. Technical Solution & Detailed Configuration – Recommend specific server specifications (e.g., NVIDIA H100 or domestic alternatives), node count, total compute power (PFLOPS), VRAM/RAM capacity, storage type & size, network bandwidth; list the software stack (PyTorch/TensorFlow, distributed training frameworks, container orchestration); and provide a full cost breakdown (hardware procurement, rack hosting, maintenance services, contingency for future scaling).

📌 3. ROI & Quantified Business Value – Estimate expected gains: reduction in model training time (e.g., from 2 weeks to 3 days), shortened feature release cycles (from 9 to 5 months), predicted NPS uplift, improvements in autonomous driving safety metrics, and the consequent boost in order conversion; calculate payback period (≤ 2 years recommended) and provide reasonable estimates for IRR and NPV.

📌 4. Risk Identification & Mitigation – Analyse key risks (compute utilisation volatility, technology roadmap shifts, supply chain delays, post‑procurement O&M cost overruns) and propose actionable countermeasures (hybrid‑cloud elastic scaling, phased procurement, backup leasing agreements with cloud providers, internal utilisation monitoring mechanisms).

📌 5. Document Structure & Tone – The proposal must include: an Executive Summary (1 page), background & strategic rationale, technical solution & budget details, ROI analysis, risk management, next steps, and a call for approval; language should be concise, professional, and persuasive for C‑level presentations; also attach descriptive data charts (text‑based trends/comparisons) and a prepared Q&A list (industry benchmarks, alternatives, and compromise options if the CPO raises objections).

✨ The final output should be a complete formal proposal (approx. 1500–2000 words) with clear structure, ready to be copied into PPT or Word; additionally, append a “Verbal talking points for the presentation” section to enhance delivery effectiveness.