Azure AI Foundry

Azure Monitor Application Insights now stores generative AI

content in a dedicated GenAIContent table (AppGenAIContent in Log

Analytics), making it possible to apply newly

available access controls to sensitive AI telemetry. Because Application

Insights powers observability in Microsoft Foundry, these protections benefit

customers using either experience. 

Since these new protections are built into Log Analytics at

the platform layer, they extend beyond just generative AI content. You can now easily

configure any table in Azure Monitor to use a more restrictive role for data

access.

Organizations can use these capabilities to help…

Preview: July 2026 · Private Preview: July 2026

Microsoft Agent Framework reaches general availability for multi-agent orchestration patterns, including the dynamic Magentic pattern that delivered 38 percent on GAIA and competitive results on AssistantBench and WebArena in published benchmarks. The GA release covers sequential pipelines, parallel collaboration, and Magentic-style coordination, all expressed in the same framework API so teams can swap patterns without rewriting agent code. Customers running complex research, document processing, and operations workflows can now build specialist agents and let the framework handle planning, hand-off, and recovery. The release ships in .NET…

Effective: June 2026

Microsoft Agent Framework reaches general availability for Agent Harness, the production runtime for agents built with the framework. Customers running pilots on the preview harness can now move the same agent code into production with backwards compatibility, SLA-backed support, and stable interfaces for tool calls, state, and telemetry. The GA release ships in .NET and Python, supports local development, container deployment, and Foundry Hosted Agents from one binary, and emits OpenTelemetry traces that show up in Foundry observability dashboards without custom wiring. Enterprise customers building internal copilots and customer-facing…

Effective: June 2026

Microsoft Foundry is now introducing Agent Framework, a unified multi-agent orchestration SDK for C# and Python. This release consolidates patterns previously spread across AutoGen and Semantic Kernel, giving developers a consistent set of abstractions for building, coordinating, and deploying agents. The Microsoft Foundry Agent Framework reduces integration complexity, helps teams avoid conflicting libraries, and streamlines long-term maintenance by providing a single, well-supported model for multi-agent workloads. This release applies to scenarios that require deterministic coordination across tasks, tools, and models, while maintaining…

Effective: April 2026

In public preview, Microsoft Foundry now offers tracing for the Microsoft Agent Framework in both Python and .NET, helping developers observe how agents reason, call tools, and interact with models during execution. This release is part of the broader hosted agents update and emits OpenTelemetry-compatible spans that you can route to Application Insights or other compatible backends, making it easier to debug failures, identify latency hotspots, and validate behavior before deploying to production. The preview is intended for evaluation and may evolve as schemas, attributes, and sampling controls are finalized ahead of general availability.…

Effective: June 2026

Microsoft Agent Framework adds the CodeAct pattern in public preview, which lets agents collapse a multi-step plan into a single executable code block rather than chaining one tool call at a time. In representative workloads, CodeAct cut end-to-end latency by roughly 50 percent and token use by more than 60 percent compared with step-by-step orchestration. The framework runs CodeAct programs inside Hyperlight, a Microsoft-built sandboxed container runtime that boots in milliseconds and isolates each execution from the host and from other agents. Developers get the productivity of letting the model write code without giving up the safety…

Effective: June 2026

Microsoft Agent Framework adds episodic procedural memory in public preview, a built-in store that lets agents remember successful task patterns and replay them on similar future requests. The framework records action sequences, tool selections, and intermediate decisions as episodes, then retrieves the closest match when a new user request looks similar. Agents become measurably faster and cheaper at recurring tasks because they reuse proven plans instead of re-deriving them from scratch on every run. The capability targets agent developers who currently bolt on custom caching layers or manual few-shot prompts to get this behavior. Episodic…

Effective: June 2026

Microsoft Agent Framework adds DevUI Agent Inspector in public preview, a local debugging surface that shows agent runs as a navigable tree of model calls, tool invocations, and intermediate state. Developers can pause, replay, and edit prompts mid-run, swap models without restarting, and inspect the full payload sent to each tool. The inspector reads the same OpenTelemetry traces that Foundry observability consumes, so what you see locally matches what shows up in production dashboards. DevUI runs as a sidecar process on a developer workstation and connects to agents built with the framework in .NET or Python. The preview is aimed at agent…

Effective: June 2026