AI 黑客松 2025
欸,這個活動到底是在做什麼呀?我們正在號召所有聰明的夥伴,一起組隊、發揮創意,開發出能翻轉我們做事方法、讓客戶滿意度爆表的 AI Agent 解決方案。
一起來施展 AI 魔法吧!
Think of AI Agents as intelligent helpers. They're computer programs that can learn from data, make choices, and take action to achieve specific goals, all without needing constant human guidance.
Example of an AI Agent: Imagine a Financial Multi-Agent that grabs data, pulls info from databases, and connects with other Agent Groups to create a super-smart system.
Example of a Non-AI Agent: A Chatbot that just spits out answers. Helpful, but not an Agent in our book.
我們正在號召所有研華夥伴,一起組隊、發揮創意,開發出能翻轉我們做事方法、讓客戶滿意度爆表的 AI Agent 點子。
一起來施展 AI 魔法吧!
報名截止
Imagine a Financial Multi-Agent that can automatically generate SQL code to retrieve data from multiple databases, seamlessly integrating with other Agent Groups to form a super-smart system. This intelligent agent not only handles data extraction and analysis but also learns from how humans collaborate and communicate across departments—enabling more effective and adaptive financial analysis.
A chatbot that just replies to questions when asked. Sure, it’s useful—but it doesn’t think for itself, take initiative, or connect with other systems to get things done. That’s why it doesn’t count as an AI Agent in this competition.
Think of AI Agents as intelligent helpers. They're computer programs that can learn from data, make choices, and take action to achieve specific goals, all without needing constant human guidance.
AI Agents are intelligent helpers—they learn from data, make decisions, and act toward goals without constant human input.
To meet this competition’s definition, an AI Agent must:
理解目標客戶的需求和偏好,自主生成不同形式的行銷內容(例如:部落格文章、白皮書、案例研究),並根據不同媒介的特性進行優化發布,監控內容效果並即時調整策略。
分析競爭對手的行銷內容與市場趨勢,協助行銷團隊生成內容主題與關鍵字建議,並依照預先設定的排程,自動將文章發佈到各種行銷媒介上。
Think of AI Agents as intelligent helpers. They're computer programs that can learn from data, make choices, and take action to achieve specific goals, all without needing constant human guidance.
具備自主學習、決策與外部環境互動的能力,能自動執行複合式任務,並提供實質的解決方案,而非被動回答。這次比賽認定的 AI Agent,至少要能做到下面兩點:
參賽者自己決定題目,從概念發想到實際做出來,最後展示成果。特別鼓勵有創意的點子,重點在於技術落地和應用創新。
經典的企業流程框架(EPF)包含了: 價值層(創造客戶價值)、賦能層(提供跨部門支持)、支撐層(提供基礎功能),依照研華特性,從 EPF 挑選出三個關鍵流程,參賽者可從中選擇任一項,設計並實做能縮短下方任一流程週期、 提升效率或滿意度的 AI Agents 解決方案
Agent 技術架構完整性 , 自主決策能力展現,開發時程與進度
業務痛點描述清晰度, 價值創造具體量化, 生產力提升明確性
架構可複製性, 使用者回饋, 跨部門應用潛力
作品將分為五大領域進行評選:業務與行銷 (Sales & Marketing)、研發工程 (R&D Engineering)、供應鏈與營運 (SCM & Operation)、智能製造 (Smart Manufacturing),以及一般管理 (General Management - CoE)。各領域獎項數量可能從缺或由評審團決議調整。
授予獎狀 1 紙,並頒發團隊獎金 新臺幣 貳拾萬元 整或成員每人 3 萬元獎金
授予獎狀 1 紙 ,並頒發團隊獎金 新臺幣 拾萬元 整或成員每人 2 萬元獎金
無論有無通過初審,與得獎分數接近但落選之作品,評審會視潛力及開發情況,斟酌給予團隊或參與成員激勵
注意:自由發想組與企業命題組獎項並無不同,差別僅在有無預先設定交件主題。
Contact Person: Paul93.Chen
Contact Person: Michael.Chuang
如果對比賽有任何問題,隨時都可以寄信到 GenAI.Center@advantech.com 詢問喔。
If you have any questions regarding the competition, please feel free to contact us at GenAI.Center@advantech.com.
All Rights Reserved | Advantech