Microsoft Adds Autonomous Agent Capabilities to Copilot

Microsoft has made a major overhaul of the Copilot artificial intelligence assistant's app structure. The previously unified app is now being divided into three sections: 'Home,' 'Code,' and the newly introduced 'Autopilot.' This change can be understood as repositioning Copilot's role from a mere conversational assistant to a tool that operates more autonomously.

The newly established Autopilot is an agent feature that operates continuously in the cloud. An agent refers to an artificial intelligence mechanism that autonomously performs pre-configured tasks without requiring users to give individual instructions. According to Microsoft, Autopilot has the ability to monitor channels in 'Microsoft Teams,' a business communication tool, and automatically complete necessary processes.

From a technical perspective, Autopilot is built on a system called OpenClaw. Its design to operate continuously on the cloud is characterized by the fact that work proceeds in the background even during times when users are not using the app. This represents a fundamental difference in operational nature compared to the previous Copilot, which was a passive tool that 'answered questions.'

The pricing model is also changing. Microsoft has indicated its intention to transition both the Autopilot and Code sections from the previous fixed-rate pricing to usage-based pricing based on consumption. Usage-based pricing is a mechanism where charges are incurred according to actual usage, and this model is commonly adopted across cloud services in general. As a result, users who rarely use the service can reduce costs, while frequent users may face increased expenses.

Behind this pricing transition lies a shift in Microsoft's business strategy. The company is understood to have adopted an 'AI assistance model' with fixed-rate pricing up to now in order to prioritize the widespread adoption of artificial intelligence services. However, the transition to usage-based pricing can be interpreted as reflecting the provision costs of AI functions more faithfully to actual usage patterns and showing a direction toward transitioning the service to a self-sustaining revenue structure.

The movement toward embedding autonomous AI agents in business software is accelerating across the industry. Features such as Autopilot, which autonomously monitor channels and complete tasks, can be seen as a concrete embodiment of a form in which artificial intelligence takes on some of the standardized communication processing and information organization work that humans have traditionally handled. How practically such features will function in actual business environments will be a point of interest going forward.

The shift from fixed-rate to usage-based pricing is also a development that has the potential to ripple across the entire artificial intelligence industry. Given the high operational costs of large language models, the pricing design strategies that companies choose will have direct implications for enterprise users' adoption decisions. Microsoft's current policy can be positioned as an example demonstrating the transition to an era where AI features are used 'only as much as needed.'