logo
2025 Microsoft Build 開發者大會:智慧代理崛起,開啟開放式 Agentic Web 新時代

2025 Microsoft Build 開發者大會:智慧代理崛起,開啟開放式 Agentic Web 新時代

Yahoo20-05-2025

AI 智慧代理(AI agents)的時代正式登場。隨著推理與知識管理技術突飛猛進,AI 模型不只變得更強大,效率也更高,展現出能用嶄新方式協助解決問題的潛力。像是 GitHub Copilot,已吸引 1,500 萬名開發者使用,其內建的代理模式和程式碼審查功能,正逐步改變開發者撰寫、檢查、部署及除錯程式碼的流程。
在企業領域,數十萬名使用者已開始透過 Microsoft 365 Copilot 進行資料研究、激盪想法與開發解方。目前已有超過 23 萬家企業組織——包含九成財星 500 大企業——使用 Copilot Studio 打造 AI agents 和自動化工作流程。
例如,富士通與 NTT Data 運用 Azure AI Foundry 協助潛在客戶排序,加快提案進度,並提供市場洞察;Stanford Health Care 則透過 healthcare agent orchestrator 建構 AI agents,進一步簡化行政作業及醫療討論會議的準備流程。
微軟在今年的 Build 開發者大會中展示了一系列針對平台、產品與基礎架構的創新,目的是讓 AI 智慧代理不只存在於概念,更實際落地在開發、研究與產業應用中。這些變革圍繞著「Agentic Web」——一種讓 AI agents 成為日常任務與決策助手的網路架構,並且是一個開放式、生態系整合的未來網路藍圖。
這樣的願景背後,有一個清楚的方向:協助開發者與組織發揮創意,驅動未來的數位轉型與創新。
在技術方面,微軟透過 GitHub、Azure AI Foundry 與 Windows 等平台,帶來多項全新工具與功能。GitHub Copilot 升級為具備智慧代理能力的程式助手,加入非同步代理功能,還新增提示管理、評估工具與控管機制;此外,GitHub Copilot Chat 在 VS Code 中也正式開放原始碼,促進更多開放式協作。
Windows AI Foundry 則成為一個從模型訓練到推論都能涵蓋的完整開發平台,支援在裝置端或雲端部署大型語言模型。Azure AI Foundry 也進一步引入模型排行榜與智能模型路由器,可根據需求自動選擇合適的模型,強化使用彈性。
在 AI agents 的建置與部署方面,微軟也推出一系列安全又容易擴展的方案,包括預設代理、客製化模組與多代理協作等功能。新登場的 Azure AI Foundry Agent Service 整合 Semantic Kernel 與 AutoGen,提供開發者友善的 SDK,同時支援模型上下文協定(Model Context Protocol, MCP)與代理互動機制(A2A)。
企業在部署 AI agents 時的資料安全與身份管理,也是本次更新的重點。Microsoft Entra Agent ID 為每個代理自動產生唯一身分識別,結合 Microsoft Purview 的資料治理機制與報告工具,有助於防堵無序擴張與潛在風險。
在 Microsoft 365 的應用層面,Copilot Tuning 功能讓企業能簡易訓練模型、建構專屬代理;同時多代理協作功能也已正式支援,可以讓不同專長的代理協同完成複雜任務,例如律師事務所就可設計專門生成法律文件的代理。
為了推動整個 AI 生態更進一步走向開放,微軟也全面支援 Model Context Protocol,並與 GitHub 共同加入其指導委員會。此外,全新開放專案 NLWeb 正式發表,未來網站可以透過 NLWeb 建立互動式對話介面,同時成為 AI agents 可探索與應用的資料端點。
最後,在 AI 加速科研應用上,微軟發表 Microsoft Discovery 平台,讓研究人員能以智慧代理技術重新設計研發流程,從藥物研發到永續發展議題都能加快探索與解決的腳步。
更多3C資訊Motorola razr 60 Ultra 實測vivo X200 Ultra 比 X200 Pro 強很多?

Orange background

Try Our AI Features

Explore what Daily8 AI can do for you:

Comments

No comments yet...

Related Articles

Google quietly released an app that lets you download and run AI models locally
Google quietly released an app that lets you download and run AI models locally

Yahoo

time2 days ago

  • Yahoo

Google quietly released an app that lets you download and run AI models locally

Last week, Google quietly released an app that lets users run a range of openly available AI models from the AI dev platform Hugging Face on their phones. Called Google AI Edge Gallery, the app is available for Android and will soon come to iOS. It allows users to find, download, and run compatible models that generate images, answer questions, write and edit code, and more. The models run offline, without needing an internet connection, tapping into supported phones' processors. AI models running in the cloud are often more powerful than their local counterparts, but they also have their downsides. Some users might be wary of sending personal or sensitive data to a remote data center, or want to have models available without needing to find a Wi-Fi or cellular connection. Google AI Edge Gallery, which Google is calling an "experimental Alpha release," can be downloaded from GitHub by following these instructions. The home screen shows shortcuts to AI tasks and capabilities like "Ask Image" and "AI Chat." Tapping on a capability pulls up a list of models suited for the task, such as Google's Gemma 3n. Google AI Edge Gallery also provides a "Prompt Lab" users can use to kick off "single-turn" tasks powered by models, like summarizing and rewriting text. The Prompt Lab comes with several task templates and configurable settings to fine-tune the models' behaviors. Your mileage may vary in terms of performance, Google warns. Modern devices with more powerful hardware will predictably run models faster, but the model size also matters. Larger models will take more time to complete a task — say, answering a question about an image — than smaller models. Google's inviting members of the developer community to give feedback on the Google AI Edge Gallery experience. The app is under an Apache 2.0 license, meaning it can be used in most contexts — commercial or otherwise — without restriction. 登入存取你的投資組合

Anthony Edwards與Kevin Durant聯手?暑假可能有機會
Anthony Edwards與Kevin Durant聯手?暑假可能有機會

Yahoo

time2 days ago

  • Yahoo

Anthony Edwards與Kevin Durant聯手?暑假可能有機會

在交易大限截止日前後,Kevin Durant驚訝地發現,鳳凰城太陽原本打算交易Bradley Beal,但與金州勇士等球隊的交易談判中,他們突然轉向交易自己。 KD(左)Edwards(右)。圖片來源:達志影像 根據ESPN的Brian Windhorst在Hoop Collective播客節目中報導,「與相關方的交談中,我清楚地意識到灰狼在交易截止日前交易Kevin Durant的認真程度。」 然而當時促成這筆交易幾乎不可能,因為兩支球隊都超過了豪華稅的上限,豪華稅對任何交易都施加了各種限制(任何球隊都不能合併球員,也不能收回比支出多1美元的金額)。 今年夏天,灰狼有3名核心球員可能,而且將會成為自由球員,他們必須做出決定,Julius Randle 與Naz Reid擁有球員選項,Nickeil Alexander-Walker是自由球員。 這些球員在7月1日不再續約,球隊將有足夠的薪資空間來重新考慮Kevin Durant的交易,問題是在季後賽第2輪,灰狼似乎找到了自己的定位和前進的道路之後,是否會想再進行一次重磅交易來改變現狀? 「我並不是說Kevin Durant最終會去明尼蘇達,如果你看看Tim Connelly(灰狼總裁),他完成了那筆意義重大的Rudy Gobert交易,他還用KAT換來了 Julius Randle 與 DiVincenzo。」 常客快艇、新興強權雷霆、金塊、灰狼、火箭、灰熊;東區龍頭騎士、衛冕軍塞爾提克、公鹿、溜馬、尼克、魔術以及大驚奇活塞、老八熱火,16支球隊齊聚一堂,就為了爭奪那座象徵年度最高榮譽的歐布萊恩金盃。精彩對戰、場場精彩敬請鎖定緯來體育台現場直播! ✎緯來體育台官方網站 【 】 ✎緯來新聞網 【 】 ✎緯來體育台官方YouTube收看NBA LIVE直播【 】 緯來體育台-2024-25NBA美國職籃大賽-賽程轉播表

Chat with Your Data: MindsDB Launches Open-Source AI Interface for Databases and Documents
Chat with Your Data: MindsDB Launches Open-Source AI Interface for Databases and Documents

Yahoo

time3 days ago

  • Yahoo

Chat with Your Data: MindsDB Launches Open-Source AI Interface for Databases and Documents

MindsDB's new AI interface enables natural language conversations with enterprise datasets—unifying structured databases and unstructured knowledge with intelligent agentic orchestration SAN FRANCISCO, May 30, 2025 /PRNewswire/ -- MindsDB, the open-source enterprise AI platform, today announced the release of their AI chat interface in MindsDB Open Source. This platform enables users to interact with their connected databases and knowledge bases using natural language—merging semantic understanding and SQL querying in a single unified experience. Drawing inspiration from MindsDB's enterprise product line, the chat interface brings the advanced conversational capabilities of intelligent agents directly into the open source offering, allowing developers, data scientists, and business users alike to "talk to their data" with no-code simplicity. Solving the "Two Data Languages" ProblemFor decades, enterprises have faced a dual-language challenge in data access: SQL is powerful for querying structured databases but requires technical fluency and schema knowledge. Semantic search tools handle unstructured content but often operate in isolation. MindsDB eliminates this divide by using a conversational interface powered by its AI Agent technology—automatically interpreting user queries and orchestrating the right mix of SQL and semantic operations behind the scenes. For example, a user can ask: "What are the common themes in support tickets about feature X, and how does that correlate with user engagement metrics?" MindsDB intelligently splits this into: A semantic query to extract themes from support tickets. A parametric SQL query to retrieve structured usage data. A unified response, delivered conversationally in the UI. Breakthrough Architecture: Built for AI Agents & Human CollaborationMindsDB is underpinned by an innovative architecture tailored for AI-native systems: Model Context Protocol (MCP): Powers standardized access to data via tools exposed by the MindsDB's Federated Query Engine, enabling seamless integration with AI agents and platforms. Agent-to-Agent (A2A) Communication: Allows the Chat Agent to coordinate with specialized SQL and semantic Agents—laying the groundwork for scalable multi-agent AI systems. Knowledge Bases: Combine vector search, embedding models, and optional reranking into a coherent semantic layer, accessible via chat and programmatic APIs. Text2SQL Translation: Automatically generates accurate SQL queries from natural language inputs for relational, AI, and federated databases. A diagram in the launch blog visualizes this multi-layered architecture—from chat input through intelligent routing and query generation to synthesized output. Key Benefits and Use CasesRespond unlocks transformative capabilities for enterprise users: Democratized Access: Enables anyone in the organization to explore enterprise data and knowledge using natural language—no data analytics expertise required. Unified Insights: Combines structured and unstructured sources into a single, coherent answer. Accelerated Development: Developers can rapidly prototype AI features and agentic workflows with reusable patterns. Open Ecosystem Integration: Fully accessible via MCP (with A2A integration coming soon), enabling third-party agents and tools to interface with MindsDB as an intelligent backend. Availability and Getting StartedThe new Chat UI is now available in beta within the latest release of MindsDB Open Source. Detailed setup instructions are available in the MindsDB documentation. About MindsDBMindsDB enables humans, AI, agents, and applications to get highly accurate answers across disparate data sources and types. Unlocking AI Search and Analytics for enterprises, MindsDB unifies petabyte-scale structured and unstructured data across diverse data sources and applications. Powered by an industry-first cognitive engine that can operate anywhere (on-prem, VPC, serverless), it empowers both humans and AI with highly informed decision-making capabilities. Media ContactZubin Tavariamedia@ View original content: SOURCE MindsDB Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data

DOWNLOAD THE APP

Get Started Now: Download the App

Ready to dive into the world of global news and events? Download our app today from your preferred app store and start exploring.
app-storeplay-store