The Microsoft Copilot Studio April 2026 updates improve visibility and governance for admins and expand workflow capabilities for managing agents.
Copilot surfaces agent status in the authoring experience, giving admins insight into each agent’s security and protection posture. Customers can identify issues such as authentication gaps or policy impacts and investigate them at the source.
Insights can be shared through the generally available Analytics Viewer role, which grants access to the Analytics page for a specific agent. Users assigned this role can view the agent’s Analytics page and metrics, and open the agent directly from the Analytics page. They cannot edit or share the agent or access topics, actions, settings, testing tools, or publishing features.
Analytics Viewer (Source: Microsoft)
The role allows analysts and stakeholders to monitor agent performance securely without affecting agent behavior. It separates responsibilities between admins who manage agents and analysts who review performance insights. Teams can access performance metrics directly, reduce the need to export and distribute analytics data manually, maintain governance over production agents, and support faster data-driven decision-making.
“The Analytics Viewer role allows us to provide meaningful performance insights to business and operational stakeholders while maintaining strict production governance. It cleanly separates operational visibility from agent configuration and publishing rights,” Mohamed Arhab, Solution Architect, City of Montreal, said.
Agent 365, now generally available, acts as a control plane for observing, governing, and securing AI agents and their interactions, including Microsoft AI agents and those from ecosystem partners, through existing admin and security workflows.
Copilot Studio customers can manage agents with shared policies, security controls, and lifecycle oversight.
Organizations can use Agent 365 to oversee AI systems acting on behalf of users through delegated access or operating independently within defined scopes and permissions.
Workflows are step-by-step automation processes in Copilot Studio that complete actions or tasks in a deterministic way. These updates expand workflow capabilities with AI-powered reasoning, centralized governance, and secure-by-design tools.
By embedding Copilot Studio agents through agent nodes, workflows can delegate reasoning, decision-making, or content generation at any step in the process. Customers can also add and configure AI actions to understand requests, route work, and generate content dynamically.
Using sample inputs to test individual steps, teams can validate behavior earlier, debug more efficiently, and refine workflows before deployment.
To scale automation without introducing new risks, workflows can connect to a broader ecosystem of tools, including model context protocol (MCP) server tools, now in preview. MCP support allows agent workflows to discover and invoke MCP-compliant tools and knowledge as workflow steps while remaining within Microsoft security, permission, and compliance boundaries. This allows workflows to execute tasks while involving users in governed review and approval processes.
“We’ve also introduced a centralized, admin-controlled environment for Workflows Agent. This makes it easier to apply data loss prevention (DLP) policies consistently and maintain visibility across automation, so workflows remain compliant by design as they scale,” Nitasha Chopra, VP & COO, Microsoft Copilot Studio, explained.
Acting on agent-generated insights often requires switching tools, recreating context, or handing off work across systems. Generally available support for apps in agents helps close that gap, reducing friction and allowing teams to review data, update records, approve requests, or create assets in place. This helps teams move faster from planning to execution.
Apps in agents bring together Microsoft and partner applications so agents can take action in systems that firms already use. Experiences are built and orchestrated in Copilot Studio, where companies can define how agents interact with apps, data, and workflows to support business processes. These generally available capabilities help teams shift agents from informational tools to operational systems.
Microsoft introduced updates to improve agent evaluation, automation, orchestration, and collaboration in Copilot Studio and Work IQ.
New evaluation features allow teams to generate test cases from analytics, simulate multi-turn interactions, and automate evaluations through APIs and connectors. Organizations can turn user conversations into test sets and run evaluations programmatically to measure agent quality at scale.
Custom metrics for outcome-based measurement allow companies to track business-specific goals such as resolution rates and conversions instead of relying only on usage data.
Microsoft also made the Work IQ API publicly available in preview, enabling developers to bring Copilot’s organizational context, memory, and signals into custom agents and workflows without managing raw data or complex integrations.
Work IQ now supports agent-to-agent communication, allowing agents to collaborate, delegate tasks, and maintain shared context between workflows.
In addition, GPT-5.5 Thinking is available in Copilot Studio early release cycle environments as GPT-5.5 Reasoning. The model is also rolling out across Microsoft 365 Copilot applications, including Copilot Chat, Word, Excel, and PowerPoint.
The expanded agent usage estimator now includes Dynamics 365 agents such as Sales Qualification Agent and Customer Service Agent. Firms can forecast Copilot credit consumption for Copilot Studio and Dynamics 365 scenarios in one place, helping them model usage more accurately and avoid unexpected costs when scaling deployments.
Microsoft said the latest admin updates are designed to reduce bottlenecks, improve decision-making, and simplify large-scale agent deployments.
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。