Microsoft made big waves this week by enhancing its Copilot ecosystem with true agentic AI capabilities. The new Microsoft 365 Copilot Chat now provides free AI assistance to frontline workers, allowing them to ask questions in natural language about company data. Meanwhile, Copilot Studio empowers companies to build custom agents like the Sales Agent, which autonomously identifies potential customers and helps close deals faster through automated outreach. IT teams are using similar tools to resolve tech tickets without human intervention, cutting downtime by up to 40%.

Google answered with major developer-focused releases. Their open-source Agent Development Kit (ADK) simplifies creating AI agents that handle complex workflows, while the Agent2Agent (A2A) protocol enables secure communication between different AI systems. Over 50 companies including Salesforce and MongoDB now use this protocol for tasks like automatically updating customer records across platforms. Google also launched Agentspace, letting workers access AI tools directly from Chrome’s search bar.

For coders, GitHub Copilot’s new agent mode represents a leap forward. When activated in VS Code, the AI can now break down feature requests into steps, edit multiple files simultaneously, run tests, and fix errors through 5-10 automatic iterations. This matches capabilities seen in top agentic coding assistants like Cursor and QodoAI, which recently gained the ability to refactor entire codebases overnight.

A detailed analysis of AI coding tools ranked GitHub Copilot Agent first for its deep Microsoft ecosystem integration, while Devin AI impressed researchers by autonomously building a working restaurant website in under an hour during tests. However, experts caution that these tools still require human oversight, as they sometimes make “confident mistakes” in complex scenarios.

The business impact continues growing as agentic AI moves beyond prototypes. A European bank reported saving $2M monthly using Microsoft’s tools to automate financial reports, while a U.S. hospital chain reduced IT costs by 30% through AI-powered system monitoring. As these technologies spread, companies are creating new roles like AI Agent Trainer to teach systems company-specific processes.