Microsoft Expands Copilot Capabilities, Rolls Out Cowork Tool

Microsoft has introduced new features for its Copilot research assistant, allowing users to employ multiple artificial intelligence models simultaneously within single workflows. This move represents the latest effort by the Big Tech organisation to enhance its AI offering and encourage greater adoption across its user base.

Credit: MICROSOFT

One new feature, named "Critique," enables Copilot's Researcher agent to draw outputs from both OpenAI's GPT and Anthropic's Claude models for each response. This contrasts with the previous method of relying on a single model for content generation.

Initially, GPT generates the primary response. Claude subsequently reviews this output for accuracy and overall quality before presenting it to the user. According to Microsoft, the company anticipates making this workflow bi-directional in the future. This would allow GPT to also review drafts produced by Claude.

But having diverse models from various vendors integrated into Copilot holds significant appeal. Nicole Herskowitz, corporate vice president of Microsoft 365 and Copilot, stated in an interview with Reuters that the organisation is advancing this integration. She indicated that customers would gain direct benefits from the models actively collaborating.

This multi-model strategy aims to accelerate user workflows, reduce instances of AI hallucinations, where systems generate factually incorrect information, and ultimately produce more dependable outputs. Herskowitz added that these improvements are expected to increase productivity and quality for customers.

The company is also launching 'Council,' a feature designed to facilitate side-by-side comparisons of responses from different AI models. These upgrades coincide with Microsoft making its new Copilot Cowork agentic AI tool more widely accessible to participants in its 'Frontier' programme. This programme grants customers early access to some of its most recent AI capabilities.

And Microsoft had previously unveiled Copilot Cowork in a testing phase earlier, leveraging Anthropic's popular Claude Cowork product. This development capitalises on the growing demand for autonomous AI agents within the industry.

The Windows maker has maintained a competitive pace, continually improving its Copilot assistant. This effort seeks to drive increased adoption amidst intense competition from rivals, including Google's Gemini, and other autonomous agents such as Claude Cowork.

Microsoft shares rose approximately 1 per cent following these announcements. However, the stock is currently on course for its poorest quarterly performance in a considerable period, with a decline of nearly 25 per cent. This trajectory reflects a broader waning of investor optimism concerning AI.

So, the company's strategic focus remains on refining its AI tools. It seeks to consolidate its position within a rapidly evolving technological landscape.

• Microsoft introduced "Critique" and "Council" features in Copilot.

• "Critique" uses OpenAI's GPT and Anthropic's Claude for multi-model response generation and review.

• The company is making its Copilot Cowork agentic AI tool more broadly available through its 'Frontier' programme.

• These enhancements aim to improve workflow efficiency, reduce AI inaccuracies, and boost output reliability.

• Microsoft faces ongoing competition from other AI developers in the autonomous agent market.

Source: Reuters