Microsoft Launches Copilot Cowork Agents

The move to "agentic AI" marks a significant evolution from earlier assistants. While generative AI creates content from prompts, agentic AI is designed to pursue goals, breaking down requests into tasks, making plans, and taking action across different applications to complete them. This leap is the culmination of decades of progress, from early chatbots like ELIZA in the 1960s to the voice-activated assistants like Siri that became mainstream in the 2010s. Microsoft's Copilot Cowork, built with technology from AI firm Anthropic, allows the assistant to work in the background across Microsoft 365 apps without constant human supervision. It can independently handle tasks like analyzing meetings, compiling research from internal and external sources, and generating documents with spreadsheets. This integration follows Anthropic's own demonstration of "Claude Cowork" in early 2026, which reportedly sent shockwaves through the software industry. To access these new capabilities, Microsoft introduced its first new enterprise license plan in roughly a decade, the Microsoft 365 E7 suite. Priced at $99 per user per month, this premium tier bundles the existing E5 plan with M365 Copilot and a new control plane called Agent 365. Agent 365, available separately for $15 per user, is designed to give IT departments the ability to manage, secure, and govern the new autonomous AI agents. Teradata's agentic AI focuses on enterprise data, enabling agents to work with both structured and unstructured information like documents, images, and audio. By integrating with tools like LangChain, its agents can orchestrate complex workflows, autonomously retrieve context from company data, and operationalize the outcomes. This allows for more sophisticated, data-driven decision-making without manual intervention. The key distinction between Anthropic's original Claude Cowork and Microsoft's implementation is the environment. Claude Cowork runs locally on a user's device, while Microsoft's Copilot Cowork operates within the company's cloud-based M365 tenant, which allows for enterprise-grade security, governance, and auditing of the agent's actions. This shift toward autonomous agents is not limited to a single AI model provider. Microsoft has stated it will employ a multi-model approach, selecting the best tool for a given task, whether from Anthropic, its primary partner OpenAI, or other providers. This strategy aims to leverage the rapid advancements across the AI industry while maintaining a secure and managed enterprise environment.