Stock Markets March 9, 2026

Microsoft integrates Anthropic tech into Copilot with new 'Copilot Cowork' for AI agents

Move brings Anthropic’s Claude capabilities into Microsoft 365 Copilot as the company seeks enterprise control and security over autonomous agents

By Leila Farooq MSFT
Microsoft integrates Anthropic tech into Copilot with new 'Copilot Cowork' for AI agents
MSFT

Microsoft announced it will incorporate Anthropic’s AI into its Copilot lineup with a new feature called Copilot Cowork, modeled on Anthropic’s Claude Cowork. The tool, designed to handle complex, multi-step tasks with limited oversight, is entering testing and will be available to early-access users later this month. Microsoft said it will also add Anthropic’s Claude Sonnet models to M365 Copilot, expanding beyond its previous reliance on OpenAI’s GPT models. Pricing details were not disclosed, though some use will be included in the $30-per-user, per-month M365 Copilot enterprise plan with additional usage available for purchase.

Key Points

  • Microsoft is introducing Copilot Cowork based on Anthropic’s Claude Cowork to bring autonomous agent functionality to M365 Copilot.
  • Anthropic’s Claude Sonnet models will be made available to M365 Copilot users; Microsoft previously relied solely on OpenAI’s GPT models.
  • Copilot Cowork is in testing and will be available to early-access users later this month; some usage is included in the $30-per-user, per-month M365 Copilot enterprise package, with extra usage purchasable.

Microsoft is expanding the capabilities of its Copilot service by incorporating technology from Anthropic, aiming to meet rising interest in autonomous AI agents. The company revealed a new offering called Copilot Cowork - a feature based on Anthropic’s Claude Cowork product - which has attracted attention for automating complex tasks such as app creation, spreadsheet construction and organizing large sets of data with relatively little human oversight.

The initiative is part of Microsoft’s effort to translate the enthusiasm for agent-style AI into enterprise-ready services. The company emphasized its long-standing relationships with corporate customers and its emphasis on security and data controls as advantages when selling these tools to organizations that want the capabilities of AI agents but are cautious about deploying them without safeguards.

Jared Spataro, who leads Microsoft’s AI-at-Work efforts, described Microsoft’s approach to data handling for Copilot Cowork, saying, "We work only in a cloud environment and we work only on behalf of the user. So you know exactly what information it (Copilot Cowork) has access to." He contrasted that with Cloud Cowork, noting that it "only works locally on the device and most companies feel 'very uncomfortable' with that, he said. 'We’re the opposite.'"

Microsoft’s announcement follows a period of volatility after Anthropic introduced new tools for Claude that raised investor concern over the potential for AI agents to disrupt traditional software businesses, contributing to a broader selloff in the software sector. Microsoft’s shares experienced a significant pullback in February, falling nearly 9% during that month.

The Copilot Cowork feature is currently in testing and Microsoft said it will be made available to early-access users later this month. The company did not provide full pricing details, but confirmed that some usage would be included in the enterprise M365 Copilot package priced at $30 per user per month, with additional usage offered as an extra purchase option.

In addition to Copilot Cowork, Microsoft is making Anthropic’s latest Claude Sonnet models available to M365 Copilot customers. Until now, Microsoft’s Copilot service had relied solely on OpenAI’s GPT models. The change further deepens Microsoft’s relationship with Anthropic at a time when investors have questioned Microsoft’s heavy reliance on OpenAI, which represents nearly 45% of Microsoft’s cloud business contract backlog.

The announcement comes as market participants and corporate buyers weigh the benefits of autonomous AI agents against operational, security and governance concerns. Microsoft is positioning its cloud-based, enterprise-focused model as the safer option for companies seeking agent functionality without ceding control over data or workflows.

Separately, some market analysis tools are actively evaluating Microsoft’s stock in light of these developments. One such tool, ProPicks AI, is described as assessing MSFT and thousands of other companies each month using over 100 financial metrics. The tool’s promotional material cites prior picks including Super Micro Computer (+185%) and AppLovin (+157%), and suggests it looks at fundamentals, momentum and valuation to identify opportunities.


Key points

  • Microsoft is launching Copilot Cowork, built on Anthropic’s Claude Cowork, to offer autonomous agent capabilities within M365 Copilot.
  • Anthropic’s Claude Sonnet models will be added to M365 Copilot, expanding Microsoft’s model mix beyond OpenAI’s GPT.
  • The new offering is entering testing and will be available to early-access users later this month; partial usage will be included in the $30-per-user, per-month M365 Copilot enterprise plan with additional usage for purchase.

Risks and uncertainties

  • Investor concerns about AI agents disrupting traditional software companies have previously triggered a selloff in the software sector, a factor that could continue to affect market sentiment.
  • Microsoft’s dependence on third-party AI providers remains under scrutiny, highlighted by the fact that OpenAI accounts for nearly 45% of Microsoft’s cloud business contract backlog.
  • Some enterprises are uneasy with agent implementations that operate locally on devices, a concern Microsoft says its cloud-based approach is designed to address.

Risks

  • Investor worries that AI agents could unsettle traditional software companies, which has contributed to a recent selloff in software stocks.
  • Questions about Microsoft’s concentration of cloud contract backlog with OpenAI - nearly 45% - raise dependency risks.
  • Enterprise hesitation about agents that run locally on devices, which Microsoft says contrasts with its cloud-only, user-authorized model.

More from Stock Markets

WuXi AppTec Shares Surge as Annual Profit More Than Doubles Mar 24, 2026 U.S. Stocks Showing Relative Resilience Amid Iran Conflict, but Risks Remain Mar 24, 2026 European New Car Registrations Tick Up in February as EV Momentum Continues Mar 24, 2026 Keymed Biosciences Set to Receive Up to $320 Million After Gilead Acquires Ouro Medicines Mar 24, 2026 Iran Fires Multiple Missile Waves at Israel as Negotiation Claims Are Disputed Mar 24, 2026