寫文章的這個時間點，我們連公司首頁都還沒有。
跟合夥人一開始的想法是：「企業怎麼導入 AI 呢？」
隨後便從已存在的關係中，找到願意嘗試的夥伴公司開始做需求訪談。我們嘗試了 Chatbot → AI 產生 EDM → AI 協助 Recruiting → AI 協助分析 SRE 部門的 alerts → AI 虛擬人應用。
Chatbot 的想法很簡單：把公司內部的行政規章、假勤規定、福利制度等文件整理後餵給 AI，用 RAG 系統做了一個可連結到公司內部員工系統的 chatbot，讓員工可以向 AI 詢問各種規定與福利。
但在實際部署後，我們發現使用率並不如預期，員工似乎更習慣直接問人資或查看公告。後續幾個嘗試與實作在有了 POC 版本後，我們也開始與關係較密切的其他潛在客戶公司做技術分享與交流。幾次下來，對方在交流過程中雖然都有表達興趣，但後續多半少有進一步的合作與落地機會。
我們逐漸理解到，成功的 AI 導入不是技術展示，而是要能真正融入使用者的工作流程，解決他們每天會遇到的具體問題。透過更多訪談與觀察，我們意識到問題的核心不在技術本身，而是我們沒有真正解決使用者的實際需求。不同規模的企業面臨的挑戰截然不同：有些需要優化既有流程，有些則希望從零開始建立 AI 應用能力。
我們發現，原先定義的「企業怎麼導入 AI 呢？」這個問題太廣泛了。
每個產業、每個部門、每間公司的情況都不同，很難有一個通用的解決方案。
到底要做什麼？
怎麼做？
簡短來說，我們目前想專注的是：
貼近客戶，快速探索
透過這段時間的探索，我們深刻體會到，創業不是一開始就要有完美的產品，而是要保持開放的心態，願意不斷調整方向。最重要的是，要真正走出去與潛在客戶對話，從他們的痛點中找到值得解決的問題。
因此，我們展開了客戶探索實驗：當有一個想法時，我們利用商業模式圖從不同面向討論、思考，並訂立第一版假設；接著分工從不同角度設計實驗 → 執行 → 驗證並修正假設，一步步修正我們的想法與作法。
我們希望能大量獲得與大家直接對話的機會與回饋。如果可以，也請支持台灣的小小新創，幫忙分享並填寫底下的問卷，大約花 5 分鐘。如果願意接受我們 15 分鐘的訪談，或對我們的領域覺得有合作機會，也請聯絡我 fansia@gmail.com。非常感謝。
任何想法都是對我們的鼓勵
如果您有招募的經驗、痛點，請幫忙留下您的想法
如果你正在思考：
「是不是要創業？」
「創業要做什麼？」
「該去哪裡找創業的點子？」
我們在這個過程中閱讀了許多好文章與書，也跟大家一起分享：
How to Get Startup Ideas（https://paulgraham.com/startupideas.html）
How to Get Small Business Ideas and Evaluate Them（https://www.the-founders-corner.com/p/how-to-get-small-business-ideas-and）
Translated by AI
The Importance of Customer Needs Exploration (Isn’t It?)
At the time of writing this article, we don’t even have a company homepage yet.
Our initial thought with my partner was: “How can enterprises adopt AI?”
We then started conducting needs interviews with partner companies willing to try, leveraging existing relationships. We experimented with Chatbot → AI-generated EDM → AI-assisted Recruiting → AI-assisted analysis of SRE department alerts → AI virtual human applications.
The Chatbot idea was straightforward: organize internal administrative regulations, leave policies, and benefit systems documents to feed into AI, creating a RAG system-based chatbot connected to the company’s internal employee system, allowing staff to inquire about various regulations and benefits.
However, after actual deployment, we found usage wasn’t as expected—employees seemed to prefer asking HR directly or checking announcements. After several attempts and implementations with POC versions, we began sharing technology and exchanging ideas with other potential client companies with whom we had closer relationships. Despite expressing interest during discussions, most had few follow-up collaboration or implementation opportunities.
We gradually understood that successful AI implementation isn’t about technology demonstration, but about truly integrating into users’ workflows to solve concrete problems they face daily. Through more interviews and observations, we realized the core issue wasn’t the technology itself, but that we weren’t truly addressing users’ actual needs. Businesses of different scales face vastly different challenges: some need to optimize existing processes, while others want to build AI application capabilities from scratch.
We discovered that our original question—”How can enterprises adopt AI?”—was too broad. Each industry, department, and company has different circumstances, making it difficult to have a universal solution.
What should we do?
How should we do it?
In short, our current focus is:
Stay close to customers, explore quickly
Through this exploration period, we deeply understand that entrepreneurship isn’t about having a perfect product from the start, but maintaining an open mindset and willingness to constantly adjust direction. Most importantly, we need to actively engage with potential customers and find worthwhile problems from their pain points.
Therefore, we launched customer exploration experiments: when we have an idea, we discuss and think from different perspectives using the business model canvas to establish our first hypothesis; then we divide tasks to design experiments → execute → validate and modify hypotheses, gradually refining our ideas and approaches.
We hope to gain numerous opportunities for direct dialogue and feedback. If possible, please support this small Taiwanese startup by sharing and filling out the questionnaire below, which takes about 5 minutes. If you’re willing to participate in our 15-minute interview or see collaboration opportunities in our field, please contact me at fansia@gmail.com. Thank you very much.
If you have recruitment experience, ideas or pain points, please share your thoughts (any idea is encouragement to us)
If you’ve encountered difficulties in job searching like me, or have ideas to try, please leave your suggestions:
If you’re considering:
“Should I start a business?”
“What business should I do?”
“Where can I find startup ideas?”
We’ve read many great articles and books throughout this process that we’d like to share:
How to Get Startup Ideas（https://paulgraham.com/startupideas.html）
How to Get Small Business Ideas and Evaluate Them（https://www.the-founders-corner.com/p/how-to-get-small-business-ideas-and）
Business Concept Monetization（https://www.eslite.com/product/1001122732681972278000?srsltid=AfmBOopm6SSf66dgAagt1mb3if9M2jiL7udAr2NJZI5CngYQ5EkTlRlL)