Agentic activity
Also written: agent activity
AI agents performing tasks, tracked by Microsoft Copilot usage metrics for custom agent activity.
Agentic activity refers to the actions taken by AI systems that are designed to operate more independently. Think of it like an AI assistant that can not only answer questions but also perform tasks for you, like scheduling a meeting or summarizing a long document without you having to explicitly tell it every single step.
Microsoft Copilot, for example, is starting to track how often custom "agents" within its system are actively working. This helps them understand which automated tasks are most useful to users and where to improve the AI's capabilities.
Agentic activity describes the operational behaviors of AI systems, particularly those designed with a degree of autonomy, in executing tasks or achieving goals. This encompasses the execution of plans, interaction with tools or environments, and decision-making processes undertaken by the agent.
Tracking agentic activity, as seen with Microsoft Copilot's usage metrics API, allows for the quantification of an agent's engagement and effectiveness. This data is crucial for understanding user adoption of agentic features, identifying areas for performance optimization, and informing future development priorities.
Microsoft Copilot is tracking custom agent usage, indicating a growing interest in understanding how AI agents perform tasks. This focus on agentic activity suggests a move towards more autonomous AI systems that can execute complex operations and be measured for their effectiveness.
- Agentic CLI customizations now in the usage metrics APIMicrosoft · GitHub Changelog: Copilot ·
- Microsoft Copilot
- Microsoft Copilot is tracking custom agent usage, allowing developers to monitor the actions of AI agents and understand their value.
- AI inference
- AI inference is the process of an AI model generating an output ba
- Prompt engineering
- Prompt engineering is about crafting inputs to guide AI beha
- Reinforcement learning
- Reinforcement learning is a training method where an AI
The concept of agentic activity stems from research in artificial intelligence and robotics, particularly concerning autonomous agents capable of perception, reasoning, and action in an environment. The term has gained prominence with the rise of large language models and their application in creating more sophisticated AI agents.
Experiment with a feature like Copilot's ability to summarize a document or draft an email. Observe how it takes your initial request and performs multiple steps to complete the task, demonstrating agentic activity. Try asking it to "create a meeting agenda for a project kickoff" and see what it generates. Notice if it asks clarifying questions or makes assumptions, which are signs of its agentic capabilities.
The Playbook: reusable how-tos, prompts, configs and tips pulled out of the news.
- PromptRole-play a student struggling with math for AI tutor simulationSimulate a student's math struggles and uncertainty for AI tutor evaluation. · 5 min
- TemplateReusable workflow templates for healthcare teamsTeams can create and share templates for common workflows like documentation or prior authorizations. · 15 min
- TemplateTemplate for building an AI tool tailored to your workA structured approach to defining and building a custom AI tool for specific business needs. · 15 min
- TemplateEvals CSV prompt list for agent skillsA CSV file template for listing prompts used in evaluating agent skills. · 20 min
- HeuristicDouble coding agent task completion with docs linksProviding coding agents with a documentation link significantly increases their task completion rate. · 2 min
Questions people ask
- What is an AI agent?
- An AI agent is a system designed to perceive its environment, make decisions, and take actions to achieve specific goals, often with a degree of autonomy.
- How is agentic activity different from just using an AI?
- Using an AI typically involves direct, step-by-step instruction for each output. Agentic activity implies the AI can perform a series of actions or a complex task with less direct supervision.
First seen on Subvolts Sep 17, 2026. Updated Sep 24, 2026. Written by an AI model from the items above and reviewed against them weekly; report an error.