# Olivia Shen
## Home
Hi, I'm
你好，我是
Olivia Shen
AI Product Commercialization
AI 产品商业化
Strategy × Go-to-Market × Operations
战略 × 市场进入 × 运营
I help AI products find product-market fit and scale. Currently building abxy.ai to help startups and enterprises turn AI capabilities into scalable business models—from pricing architecture to revenue operations and GTM execution.
我帮助 AI 产品找到产品市场契合点并实现规模化。目前正在建设 abxy.ai，帮助初创企业和大型企业将 AI 能力转化为可扩展的商业模式——从定价架构到收入运营和市场进入执行。
View My Work
查看我的作品
| Platform | URL | Icon |
| --- | --- | --- |
| LinkedIn | https://www.linkedin.com/in/ruofan-shen | :icon-linkedin: |
| Substack | https://notoliva.substack.com/ | :icon-pen-line: |
| GitHub | https://github.com/oliviashenrf | :icon-github: |
| Email | mailto:shenr20wlu@gmail.com | :icon-mail: |
## Products
Products I Built
我构建的产品
Tools and prototypes built through vibe coding and AI
通过氛围编程和 AI 构建的工具和原型
Products I Built
我构建的产品
Screenshot coming soon
截图即将上传
| Name | Description | Tech Stack | Icon | Link | Status | Brand Color | Screenshot |
| --- | --- | --- | --- | --- | --- | --- | --- |
| AI-Optimized ROI Calculator | Built for startups to rapidly customize GTM financial models. Enables quick decision-making and improved forecast accuracy for investment returns. Used by multiple early-stage companies for investor presentations. | AI-assisted coding, JavaScript, Financial modeling | :icon-calculator: | https://roi-forge-399009678734.us-west1.run.app/ | Live | #4E46DC |  |
| GTM Copilot | An agent-native go-to-market system that encodes your sales motion, competitive intelligence, and institutional knowledge directly into AI. It operates as an embedded teammate — executing discovery, research, positioning, and follow-ups with context, consistency, and speed — so lean teams can move faster than headcount ever could. | AI/LLM integration, GTM Motion, Chatbot, Lightweight deployment | :icon-bot-message-square: | https://gtm-operator-399009678734.us-west1.run.app/ | Live | #5AC066 |  |
| Relay - P2P Furniture Marketplace | Community-driven furniture marketplace with anti-flipping verification. Connects verified neighbors for sustainable furniture exchange, reducing waste while building local community trust. Features secure authentication and listing management. | Full-stack web app, User authentication, Community verification | :icon-truck: | https://relay-p2p-furniture-marketplace-399009678734.us-west1.run.app/#/ | Live | #0F766E |  |
| Pulse - Body Transformation Tracker | Personal health and fitness tracking app that helps users monitor body changes over time. Features clean, minimalist design with weight and height tracking, progress visualization, and motivational elements focused on discipline, persistence, and transformation. | Full-stack web app, Data visualization, Progress tracking | :icon-activity: | https://pulse-399009678734.us-west1.run.app/ | Beta | #0EA5E9 | |
## Workflows
Workflows I Defined
我定义的工作流
GTM processes, tools, and dashboards I've built and operationalized across sales, marketing, and revenue teams.
我在销售、市场营销和收入团队中构建并落地的 GTM 流程、工具和仪表盘。
Operational Methodology
实战方法论
More workflows being documented · Stay tuned
更多工作流正在整理中 · 敬请期待
Why we built it
为什么要构建它
Traditional battlecards are static and aren't customized to a specific deal. Competitive intelligence is also bottlenecked by Product Marketing bandwidth to conduct research and analysis. BattleBot solves both: we scale human insight by automating pipelines of competitive intel and win-loss analysis, then make that insight accessible and dynamic through a conversational interface.
传统竞争情报卡是静态的，无法针对具体交易定制。竞争情报的生产也受限于产品市场营销团队的带宽。BattleBot 同时解决这两个问题：我们通过自动化竞争情报和赢/输分析管道来放大人类洞察，再通过对话界面让这些洞察变得可访问且动态化。
The workflow
工作流程
| Step | Step (Chinese) | Detail | Detail (Chinese) | Icon |
| --- | --- | --- | --- | --- |
| Daily Research Pipelines | 每日研究管道 | Back-end pipelines run daily across many dimensions — news, sales calls, CRM opportunities, and more. | 后端管道每天跨多个维度运行，包括新闻、销售通话、CRM 商机等。 | :icon-database: |
| AI Analysis + Human-in-the-Loop | AI 分析 + 人工审核 | AI conducts a first layer of analysis. A human reviews and approves before insights enter the knowledge base. | AI 进行第一层分析，人工审核批准后洞察才进入知识库。 | :icon-user-check: |
| Win/Loss Intelligence | 赢/输情报 | Retool + Gemini power granular win/loss analyses and first-level summaries from deal data. | Retool + Gemini 驱动基于交易数据的细粒度赢/输分析和一级摘要。 | :icon-trophy: |
| Conversational Assistant for Sellers | 面向销售的对话式助手 | Knowledge is exposed to an assistant built in Dust (Claude Sonnet under the hood). Sellers get contextual competitive guidance in the flow of work — a prompt away. | 知识通过 Dust 构建的助手（底层使用 Claude Sonnet）呈现。销售人员在工作流中即可获得情境化的竞争指导，一句话搞定。 | :icon-message-circle: |
Tech stack
技术栈
| Tool | Role | Role (Chinese) | Icon |
| --- | --- | --- | --- |
| Retool | Workflow orchestration & dashboards | 工作流编排与仪表盘 | :icon-layout-dashboard: |
| Gemini | Research pipelines & win/loss analysis | 研究管道与赢/输分析 | :icon-sparkles: |
| Dust | Conversational assistant platform | 对话式助手平台 | :icon-bot: |
| Claude Sonnet | LLM powering the seller-facing assistant | 驱动面向销售助手的大语言模型 | :icon-cpu: |
Why we built it
为什么要构建它
Sales reps spend more time updating CRM than selling, and managers lack real-time visibility into deal health. Pipeline Intelligence Copilot connects directly to Salesforce to surface at-risk deals, score engagement, and generate AI-powered next-step recommendations — so reps focus on closing, not reporting.
销售代表花在更新 CRM 上的时间比销售更多，而管理者缺乏对交易健康状况的实时可见性。Pipeline Intelligence Copilot 直接连接 Salesforce，实时发现高风险交易、评估参与度，并生成 AI 驱动的下一步建议——让销售专注于成交，而不是填报告。
The workflow
工作流程
| Step | Step (Chinese) | Detail | Detail (Chinese) | Icon |
| --- | --- | --- | --- | --- |
| Salesforce Sync | Salesforce 同步 | Live CRM data pulled automatically — deal stages, engagement scores, ICP match, and close dates. | 自动拉取实时 CRM 数据，包括交易阶段、参与度评分、ICP 匹配度和预计关闭日期。 | :icon-database: |
| Pipeline & Risk Scoring | 管道与风险评分 | Gemini 1.5 analyzes deal signals to score risk, flag stalled deals, and calculate win probability. | Gemini 1.5 分析交易信号，评估风险、标记停滞交易并计算赢率。 | :icon-shield-alert: |
| Deal Intelligence Sidebar | 交易情报侧边栏 | Click any deal to see a full risk assessment, risk factors breakdown, and AI-generated next steps. | 点击任意交易，查看完整风险评估、风险因素分析和 AI 生成的下一步行动建议。 | :icon-panel-right: |
| Salesforce Entry Draft | Salesforce 录入草稿 | AI auto-drafts the CRM update entry for each deal — one click to push back to Salesforce. | AI 自动起草每笔交易的 CRM 更新录入，一键同步回 Salesforce。 | :icon-file-pen: |
Tech stack
技术栈
| Tool | Role | Role (Chinese) | Icon |
| --- | --- | --- | --- |
| Salesforce | CRM data source | CRM 数据源 | :icon-cloud: |
| Gemini 1.5 | Risk scoring & deal intelligence | 风险评分与交易情报 | :icon-sparkles: |
| Retool | Dashboard & workflow orchestration | 仪表盘与工作流编排 | :icon-layout-dashboard: |
Why we built it
为什么要构建它
Sales teams working with complex robotics RFPs were spending 4+ hours manually analyzing 50-page documents packed with technical specs and contractual requirements. I built an RFP analysis assistant powered by Gemini, but the real work was iterative prompt engineering: testing against real RFPs, gathering feedback from sales engineers, and structuring templates around specific solution capabilities — welding, material handling, inspection. The result reduced analysis time from 4 hours to 30 minutes and improved bid quality with more targeted responses.
机器人领域的销售团队每次手动分析复杂 RFP 要花 4 小时以上——这些文件往往超过 50 页，充满了技术规格和合同要求。我基于 Gemini 构建了一套 RFP 分析助手，核心在于持续迭代的提示工程：用真实 RFP 测试、收集销售工程师反馈，并针对具体解决方案能力（焊接、物料搬运、质检）结构化提示模板。最终将分析时间从 4 小时压缩至 30 分钟，显著提升了投标质量。
The workflow
工作流程
| Step | Step (Chinese) | Detail | Detail (Chinese) | Icon |
| --- | --- | --- | --- | --- |
| Paste RFP & Select Mode | 粘贴 RFP 并选择分析模式 | Rep pastes the RFP document text and selects the robotics domain (welding, material handling, inspection) to ground the analysis. | 销售代表粘贴 RFP 文档内容，选择机器人应用领域（焊接、物料搬运、质检）以锚定分析上下文。 | :icon-file-text: |
| Choose Analysis Objective | 选择分析目标 | Select from four templates: Technical Feasibility, Risk Identification, Competitive Positioning, or Requirement Extraction — each backed by a purpose-built prompt. | 从四个模板中选择：技术可行性、风险识别、竞争定位或需求提取——每个模板都有专门构建的提示。 | :icon-list-checks: |
| Gemini AI Analysis | Gemini AI 分析 | Gemini runs the structured prompt against the RFP, returning feasibility scores, red flags, competitive differentiators, or a parsed requirements list. | Gemini 将结构化提示应用于 RFP，输出可行性评分、风险红旗、竞争差异点或解析后的需求清单。 | :icon-sparkles: |
| Feedback Loop & Refinement | 反馈循环与持续优化 | Reps rate AI outputs directly in the tool. Feedback is used to continuously refine the prompt library across robotics applications. | 销售代表在工具内直接对 AI 输出评分，反馈用于持续优化覆盖各机器人应用场景的提示库。 | :icon-refresh-cw: |
Tech stack
技术栈
| Tool | Role | Role (Chinese) | Icon |
| --- | --- | --- | --- |
| Gemini | RFP analysis & structured extraction | RFP 分析与结构化提取 | :icon-sparkles: |
| Prompt Library | 4 purpose-built templates per robotics domain | 4 套针对机器人领域的专用提示模板 | :icon-library: |
| Retool | UI, workflow orchestration & feedback loop | 界面、工作流编排与反馈循环 | :icon-layout-dashboard: |
Why we built it
为什么要构建它
When note-retention drops, teams often learn about it only after several days of mounting provider frustration — by then, clinicians have spent hours rewriting drafts. Provider edits arrive as raw JSON diffs scattered across logs, making root-cause analysis painfully slow. This dashboard ingests edit-event streams in near real time, surfaces retention percentage by site, provider, and note type, and triggers color-coded alerts before frustration escalates to Customer Success.
当笔记保留率下降时，团队往往在数天后才察觉，此时医生已花费数小时重写草稿。Provider 的编辑以原始 JSON diff 形式分散在各日志中，根因分析耗时极长。该仪表盘近实时摄取编辑事件流，按站点、Provider 和笔记类型展示保留率，并在问题升级至客户成功团队之前触发色码预警。
The workflow
工作流程
| Step | Step (Chinese) | Detail | Detail (Chinese) | Icon |
| --- | --- | --- | --- | --- |
| Edit-Event Stream Ingestion | 编辑事件流摄取 | Real-time pipeline ingests note edit events — AI drafts, provider edits, final signatures — tagged by site, provider, and note type. | 实时管道摄取笔记编辑事件——AI 草稿、Provider 编辑、最终签名——并按站点、Provider 和笔记类型打标。 | :icon-zap: |
| Funnel & Retention Scoring | 漏斗与保留率评分 | Each note flows through a 3-stage funnel. Retention ratio is calculated per note, then aggregated by site, provider, and note type with risk-tier classification (At Risk / Good). | 每条笔记流经三阶段漏斗，逐条计算保留率，再按站点、Provider 和笔记类型聚合，并分类为风险等级（高风险 / 良好）。 | :icon-funnel: |
| Color-Coded Alerts & Drill-Downs | 色码预警与下钻分析 | Metrics breaching thresholds trigger color-coded alerts. Customer Success can one-click drill into any site or provider to see the full note-level breakdown before escalation. | 超过阈值的指标触发色码预警。客户成功团队可一键下钻到任意站点或 Provider，在问题升级前查看完整的笔记级明细。 | :icon-alert-triangle: |
| Conversion Trend Visualization | 转化趋势可视化 | Cumulative conversion trend chart shows AI drafts → provider edits → final signatures over time, segmented by AI cohort, to identify template drift and sudden drop-offs. | 累积转化趋势图展示 AI 草稿→Provider 编辑→最终签名的时序变化，按 AI cohort 分段，用于识别模板漂移和突发下滑。 | :icon-trending-up: |
Tech stack
技术栈
| Tool | Role | Role (Chinese) | Icon |
| --- | --- | --- | --- |
| Retool | Dashboard, filters & drill-down UI | 仪表盘、筛选器与下钻界面 | :icon-layout-dashboard: |
| Edit-Event Stream | Real-time note lifecycle ingestion | 实时笔记生命周期摄取 | :icon-activity: |
| Risk Threshold Engine | Color-coded alerts by site & provider | 按站点和 Provider 的色码预警引擎 | :icon-shield-alert: |
## Reflections
Reflections
思考
Thoughts on AI, strategy, product, and the future of technology. Subscribe to get updates on new reflections.
关于 AI、战略、产品和技术未来的思考。订阅以获取新文章更新。
View all on Substack
在 Substack 查看全部
https://notoliva.substack.com/
| Title | Excerpt | Topic | Read Time | URL |
| --- | --- | --- | --- | --- |
| How I Think About Revenue Growth in AI Products (It's Not What You'd Expect) | Why usage rises but revenue stays flat. Treating monetization as product design, not a sales afterthought. Covers AI economics, activation optimization, and staged rollout strategies. | Revenue & Pricing | 8 min read | https://notoliva.substack.com/p/how-i-think-about-revenue-growth |
| The Rise of Vibe Coding: How I Stopped Writing Code and Started Orchestrating AI | From manual debugging to high-level orchestration. Choosing the right models, multi-layer agent systems, cost optimization, and the divergence in developer career paths. | SaaS & AI | 7 min read | https://notoliva.substack.com/p/the-rise-of-vibe-coding-how-i-stopped |
| The Death of SaaS as We Know It: 5 Counter-Intuitive Truths from the AI Frontier | Software economics are collapsing. AI companies have service-like margins, database moats are vanishing, and seat-based pricing is dying. New growth patterns, context as moat, and hybrid pricing. | SaaS & AI | 10 min read | https://notoliva.substack.com/p/the-death-of-saas-as-we-know-it-5 |
| Why SaaS Growth Can't Just Be About Acquiring New Customers | Why companies hit a ceiling at $2-5M ARR. Four-engine growth framework, expansion revenue's 7-9x better LTV/CAC, value metric pricing, and scaling strategies from $1M to $100M+. These engines run in parallel, not sequentially. | Growth Strategy | 9 min read | https://notoliva.substack.com/p/why-saas-growth-cant-just-be-about |
## AI Podcast
AI Podcast
AI 播客
An AI-powered podcast exploring the intersection of technology, strategy, and innovation—built with NotebookLM.
一个由 AI 驱动的播客，探索技术、战略和创新的交叉点——由 NotebookLM 构建。
https://youtu.be/RyZV5pTJebA
https://www.youtube.com/embed/RyZV5pTJebA
Watch on YouTube
在 YouTube 观看
## Learning Resources
Learning Resources
学习资源
Resources and tools I've discovered and relied on throughout my journey in AI, strategy, and product development
在我的 AI、战略和产品开发之旅中发现并依赖的资源和工具
| ID | Category | Description | Icon |
| --- | --- | --- | --- |
| 1 | Tools I Use | Development platforms and AI tools I rely on for building and experimenting | :icon-wrench: |
| 2 | Frameworks I Study | Business frameworks, pricing models, and strategic insights that shaped my thinking | :icon-trending-up: |
| 3 | Content I Follow | Newsletters and podcasts I read/listen to stay updated on AI, tech, and business | :icon-podcast: |
| 4 | Products I Admire | AI products with exceptional UX and growth strategies that inspire me | :icon-sparkles: |
| Category ID | Title | Description | URL | Icon |
| --- | --- | --- | --- | --- |
| 1 | Google AI Studio | Google's platform for experimenting with generative AI models, prompt engineering, and API integration | https://aistudio.google.com/ | :icon-sparkles: |
| 1 | Vercel | Frontend cloud platform for deploying and scaling web applications with seamless Git integration | https://vercel.com/ | :icon-triangle: |
| 1 | Cursor | AI-powered code editor designed for pair programming with AI, built on VS Code | https://cursor.sh/ | :icon-mouse-pointer-2: |
| 2 | The New Business of AI | a16z's deep dive into how AI companies differ from traditional software - covering margins, scaling, and defensibility | https://a16z.com/the-new-business-of-ai-and-how-its-different-from-traditional-software/ | :icon-lightbulb: |
| 2 | Inside OpenAI with Logan Kilpatrick | Lenny's Podcast episode: Head of Developer Relations shares insights on building with AI and OpenAI's strategy | https://fireflies.ai/blog/inside-openai-logan-kilpatrick-head-of-developer-relations/ | :icon-brain: |
| 2 | The State of AI 2025 | BVP's comprehensive annual report on AI market trends, valuations, and growth trajectories | https://www.bvp.com/atlas/the-state-of-ai-2025 | :icon-bar-chart-3: |
| 2 | Aggregation Theory | Ben Thompson's foundational framework for understanding how platforms create value by aggregating users and suppliers | https://stratechery.com/2015/aggregation-theory/ | :icon-network: |
| 2 | The State of Usage-Based Pricing | Kyle Poyar's comprehensive analysis on SaaS pricing evolution and hybrid monetization models | https://www.growthunhinged.com/p/the-state-of-usage-based-pricing | :icon-dollar-sign: |
| 2 | Cloud 100 Benchmarks Report | BVP's annual ranking and analysis of top private cloud companies with key performance metrics | https://www.bvp.com/atlas/the-cloud-100-benchmarks-report | :icon-cloud: |
| 2 | Battery Ventures Cloud Software Spending Report | Data-driven insights on enterprise cloud software spending patterns and market trends | https://www.battery.com/wp-content/uploads/2023/03/Battery-Ventures-State-of-Cloud-Software-Spending-Report-March-2023.pdf | :icon-file-text: |
| 3 | Growth Unhinged | Kyle Poyar's newsletter covering product-led growth, pricing strategy, and go-to-market excellence | https://www.growthunhinged.com/ | :icon-mail: |
| 3 | Benedict's Newsletter | Benedict Evans' weekly newsletter on technology trends, media evolution, and the future of tech | https://www.ben-evans.com/newsletter | :icon-newspaper: |
| 3 | Rundown AI | Daily AI newsletter covering the latest developments, tools, and trends in artificial intelligence | https://www.therundown.ai/ | :icon-zap: |
| 3 | Founders Podcast | David Senra explores the lives and strategies of history's greatest entrepreneurs and business builders | https://www.founderspodcast.com/ | :icon-headphones: |
| 3 | Acquired | Ben Gilbert and David Rosenthal dive deep into the stories and strategies of great companies | https://www.acquired.fm/ | :icon-podcast: |
| 3 | No Priors | AI podcast featuring in-depth conversations with leading researchers and builders in artificial intelligence | https://www.nopriors.com/ | :icon-radio: |
| 4 | Notion AI | AI-powered workspace that seamlessly integrates writing assistance, knowledge management, and collaboration | https://www.notion.so/product/ai | :icon-notebook-pen: |
| 4 | Jasper | AI copilot for marketing teams - content creation, brand voice consistency, and campaign workflows | https://www.jasper.ai/ | :icon-pen-tool: |
| 4 | Copy.ai | GTM AI platform that automates research, personalization, and outbound workflows for sales teams | https://www.copy.ai/ | :icon-message-square: |
| 4 | Superhuman | AI-powered email client with instant triage, auto-generated replies, and blazing-fast keyboard shortcuts | https://superhuman.com/ | :icon-zap: |
| 4 | Loom | Video messaging platform with AI-powered transcription, summaries, and async collaboration features | https://www.loom.com/ | :icon-video: |
| 4 | faces.app | AI-powered platform for building beautiful, modular web experiences through natural conversation | https://faces.app/ | :icon-sparkles: |
## About Me
About Me
关于我
I'm a strategist and builder focused on AI product commercialization. My journey spans from analyzing $30bn+ equity strategies at Russell Investments to driving revenue operations at Intrinsic AI (Google X).
In 2024, I founded **abxy.ai** to help AI companies bridge the gap between technology and sustainable business models. We focus on pricing architecture, packaging strategy, and go-to-market execution.
I believe the best way to understand a problem is to build something that solves it. Through 'vibe coding', I've shipped products ranging from GTM financial tools to marketplaces and health trackers—functional solutions that solve real problems.
I studied liberal arts at Washington and Lee and engineering at Duke—a combination that shaped how I think: strategic frameworks meet technical execution and a builder's mindset. Based in Seattle, deep in the tech ecosystem. Always open to connecting with founders in the AI space.
我是一名专注于 AI 产品商业化的战略家和构建者。我的经历从在 Russell Investments 分析 300 亿美元以上的股票策略，到在 Intrinsic AI（Google X）推动收入运营。
2024 年，我创立了 **abxy.ai**，帮助 AI 公司在技术和可持续商业模式之间架起桥梁。我们专注于定价架构、打包策略和市场进入执行。
我相信理解问题的最好方式是构建解决方案。我已经发布了从 GTM 金融工具到市场和健康追踪器的产品。
我在华盛顿与李大学学习文科，在杜克大学学习工程。这种组合塑造了我的思维方式：战略框架与技术执行以及构建者思维的结合。我位于西雅图，深度融入科技生态系统。始终欢迎与 AI 领域的创始人联系。
## Contact
Get in Touch
联系我
Have a project in mind or want to chat about AI strategy? I'd love to hear from you.
有项目想法或想聊聊 AI 战略？我很乐意与您交流。
Name
Email
Message
Your Name
your.email@example.com
Tell me about your project or question...
Send Message
Thanks for reaching out! I'll get back to you soon.