跳到主要内容Skip to main content
← Automation

Microsoft Copilot StudioMicrosoft Copilot Studio

Microsoft
企业低代码Enterprise low-code 云托管Cloud-hosted PowerPlatformPowerPlatform M365M365 agentagent partial lifecycle: active 核验Checked 2026-10-30

Copilot Studio 是 Microsoft 生态里的企业低代码 agent 平台——它的「能不能用」很大程度取决于 你在不在 Power Platform/Dataverse 生态里。企业治理(身份、数据驻留)随 Microsoft 365 体系现成提供。

Copilot Studio is Microsoft's enterprise low-code agent platform — whether it is usable largely depends on whether you are inside the Power Platform/Dataverse ecosystem. Enterprise governance (identity, data residency) comes ready-made with the Microsoft 365 system.

适合与不适合When to use it, when not to

适合:M365/Power Platform 存量企业、要现成企业治理、agent 要在 Teams/官网/M365 内分发。 不适合:不在 Microsoft 生态(生态外价值骤减);要自托管(无此形态)。

A fit for: M365/Power Platform incumbents, ready-made enterprise governance, agents distributed inside Teams/website/M365. Not a fit for teams outside the Microsoft ecosystem (value drops sharply); self-hosting (no such form).

中文生态China availability 国产 / 国内可直接用 / 备案状态 · 快照domestic / direct access / CAC filing · snapshot 2026-10-21

出品属地Origin
国外出品foreign-made Microsoft(美国)。Microsoft (US).
国内使用Direct access
直连未核验direct unknown 产品页直抓 403(CF 拦截)、文档线 200——整体直连性标 unknown。product page 403 (Cloudflare), docs 200 — overall directness marked unknown.
备案状态CAC filing
未见于备案名单not in filing list:未见于网信办公示名单(2026-10-21 检索);该产品未面向大陆公众提供服务,备案不适用。not in the CAC public registry (searched 2026-10-21); the product does not serve the mainland public, so filing does not apply.

备案信息以网信办《生成式人工智能服务已备案信息》公示名单为准(按批次滚动更新,本站只抄编号不解读);「大陆可直连」是本站从大陆网络直连实测的结果,不是厂商承诺。GitHub 类站点当日实测有波动,口径写在依据里。Filing status is quoted from the CAC public registry (updated batch by batch; this site copies the number and does not interpret it). "Direct access" is this site's own reachability measurement from a mainland network, not a vendor promise. GitHub-hosted sites fluctuated on the measurement day — the note says so.

固定坐标系Fixed coordinate system

8 个维度dimensions, 与同分区其他对象逐项可比each comparable item by item against others in the same section.

模型与开放条件Model & openness

官方文档线(learn.microsoft.com/microsoft-copilot-studio,2026-10-21 直抓 200)覆盖模型与生成式答案配置; 具体可选模型清单未逐条核验。平台与 Microsoft 生态(Dataverse/Power Platform)深度绑定。

The official docs line (learn.microsoft.com/microsoft-copilot-studio, fetched directly 200 on 2026-10-21) covers model and generative-answers configuration; the selectable-model list has not been verified item by item. The platform is deeply bound to the Microsoft ecosystem (Dataverse/Power Platform).

运行位置Where it runs

纯云托管:agent 跑在 Microsoft 365 / Power Platform 云上;无自托管形态。

Pure cloud hosting: agents run on the Microsoft 365 / Power Platform cloud; no self-hosted form.

本地文件Local files

知识源接 SharePoint/文件/Dataverse 等平台内数据;不承载本机文件运行时。

Knowledge sources connect to in-platform data such as SharePoint/files/Dataverse; it hosts no local-file runtime.

关机后的任务Tasks after shutdown

云托管常驻是默认形态;触发与自动化经 Power Automate 衔接(官方文档线)。

Cloud-hosted residency is the default form; triggers and automation connect via Power Automate (official docs line).

工具与扩展Tools & extensions

官方文档列主题/动作/连接器体系(Power Platform 连接器生态);MCP 支持状态未逐条核验。

Official docs list the topics/actions/connectors system (Power Platform connector ecosystem); MCP support status not verified item by item.

上下文与记忆Context & memory

对话主题(topics)与变量是官方上下文组织方式;跨会话记忆机制未逐条核验。

Conversation topics and variables are the official context-organization scheme; cross-session memory mechanisms not verified item by item.

权限与限制Permissions & limits

企业级身份/治理随 Microsoft 365 体系(Entra ID 等);数据驻留与合规随 Microsoft 商业条款。

Enterprise identity/governance follows the Microsoft 365 system (Entra ID etc.); data residency and compliance follow Microsoft commercial terms.

适合什么任务Best for

适合:Microsoft 365/Power Platform 存量企业、要企业治理现成、要 M365 生态内分发。 不适合:不在 Microsoft 生态;要自托管。

Fits: Microsoft 365/Power Platform-installed enterprises, governance ready-made, distribution inside the M365 ecosystem. Does not fit: outside the Microsoft ecosystem; self-hosting.

头号误解Top misconception

  • 把它当独立聊天机器人工具——它的价值绑定在 Power Platform/Dataverse 生态里Treating it as a standalone chatbot tool — its value is bound to the Power Platform/Dataverse ecosystem
  • 计费按消息/容量计量——本站按 D9 不采价格数字,看官方 pricing 页Message/measurement-based billing — this site stores no price numbers (D9); see the official pricing page
  • 官网产品页直抓 403(CF 拦截)——核验走 learn.microsoft.com 文档线The product page returns 403 to direct fetch (Cloudflare) — verification goes through the learn.microsoft.com docs line

未知项清单Known unknowns

  • 可选模型清单与切换粒度(未逐条核验)
  • MCP 支持状态(未核验到官方文档)
  • 跨会话记忆机制(未核验到官方文档)
  • The selectable-model list and switching granularity (not verified item by item)
  • MCP support status (no official documentation found)
  • Cross-session memory mechanism (no official documentation found)

证据来源Evidence sources

判断可回到以下一手源复核。本站核验日Every judgement can be rechecked against the primary sources below. Verified on 2026-10-30,内容更新日content updated on 2026-10-21。

类型Type名称Name链接Link
docsMicrosoft Copilot Studio · 官方文档入口(2026-10-21 直抓 200)Microsoft Copilot Studio · official docs entry (fetched directly 200 on 2026-10-21)https://learn.microsoft.com/en-us/microsoft-copilot-studio/
docsCopilot Studio · 官方产品页(直抓 403=CF 拦截,浏览器可达)Copilot Studio · official product page (direct fetch 403 = Cloudflare block; reachable via browser)https://www.microsoft.com/en-us/microsoft-copilot/microsoft-copilot-studio
changelogMicrosoft Learn · Copilot Studio 发布记录(changelog 入口)Microsoft Learn · Copilot Studio release notes (changelog entry point)https://learn.microsoft.com/en-us/microsoft-copilot-studio/whats-new

实测记录Measured runs

本档未跑工作流搭建实测;结论只覆盖平台能力与权限边界。

This site has not completed measured runs. The closest test protocol is the code-side protocols/framework/_protocol.md (the tool-calling and credential-boundary sections are reusable; no workflow-platform-specific protocol exists on this site yet).

相关条目Related entries

本页由This page is generated from the ai-handbook 数据层生成(CC BY 4.0)。data layer (CC BY 4.0). 方法论与坐标系定义见仓库内For the methodology and coordinate definitions, see METHODOLOGY.md.