Key Moments
- Microsoft introduced new Copilot features that let users tap multiple AI models within a single research workflow.
- The new “Critique” capability combines OpenAI’s GPT and Anthropic’s Claude, with one model generating responses and the other reviewing them.
- Microsoft is broadening early access to its Copilot Cowork agentic AI tool through its Frontier program.
New Multi-Model Capabilities in Copilot
Microsoft on Monday announced a set of upgrades to its Copilot research assistant designed to let users work with several artificial intelligence models at once in a single workflow. The company framed the move as part of its ongoing effort to enhance its AI product suite and encourage broader adoption.
The headline feature, called “Critique,” equips Copilot’s Researcher agent to draw on outputs from both OpenAI’s GPT and Anthropic’s Claude models for each response, instead of depending on just one model.
Under this setup, GPT is responsible for generating the initial answer, while Claude evaluates that response for accuracy and overall quality before it is delivered to the user, Microsoft said. The company noted it plans to eventually make the process bi-directional so that GPT can also review drafts produced by Claude.
Intended Benefits for Accuracy and Productivity
Nicole Herskowitz, corporate vice president of Microsoft 365 and Copilot, said in an interview with Reuters that having “various different models from different vendors in Copilot is highly attractive – but we’re taking this to the next level, where customers actually get the benefits of the models working together.”
According to Herskowitz, this multi-model approach is expected to accelerate user workflows, help curb AI hallucinations – where systems generate false information – and deliver more dependable results, ultimately improving both productivity and output quality for customers.
Model Comparison Via ‘model Council’
Alongside Critique, Microsoft is rolling out a feature called ‘model Council’. This tool will let users compare responses from different AI models side-by-side, giving them a direct view of how the models perform on the same prompt.
| Feature | Function | Models Involved |
|---|---|---|
| Critique | Generates a response with one model and reviews it with another to improve accuracy and quality | OpenAI’s GPT and Anthropic’s Claude |
| model Council | Displays multiple model outputs side-by-side for comparison | Multiple AI models, including those from different vendors |
Broader Access to Copilot Cowork
The new capabilities are being introduced as Microsoft expands access to its Copilot Cowork agentic AI tool to more customers in its ‘Frontier’ program, which offers early access to some of the company’s newest AI features.
Microsoft had unveiled Copilot Cowork earlier this month in testing mode. The tool is based on Anthropic’s popular Claude Cowork product and aims to tap into rising interest in autonomous AI agents.
Competitive Landscape in AI Assistants
The company, referred to as the Windows maker, has been pushing to upgrade its Copilot assistant to strengthen user adoption amid fierce competition. Key rivals include Google’s Gemini and autonomous agents such as Claude Cowork.
TradingPedia.com is a financial media specialized in providing daily news and education covering Forex, equities and commodities. Our academies for traders cover 
