Microsoft adds Claude tools to Copilot for sharper outputs
Microsoft has refreshed the Copilot interface in web and Teams. Users now see simple buttons for Learn, Find, Summarize and Suggest tasks. This targets the low adoption rates under 10 percent in organisations that bought licences. Two new preview tools combine OpenAI's GPT models with Anthropic's Claude. The Critique Model refines outputs by having Claude check GPT work. Model Council runs both models separately then summarises where they agree. Other updates include Outlook calendar rules to auto-handle meeting invites and mind maps from Copilot Notebooks. These build on 300 million monthly active users across apps like Word, Excel and Teams.
Copilot users stuck to manual work or ChatGPT because single-model outputs often felt generic or unreliable, especially with company data. Prompts in Outlook or Teams produced summaries that missed context, eroding trust amid governance worries. Multi-model previews now pit GPT against Claude for critiques and consensus reports, yielding sharper results tied to your M365 files. Calendar automation cuts scheduling drudgery, while refreshed UI lowers the entry barrier – a direct fix for the 'tried once, quit' cycle.
Analysis
This Claude boost finally makes Copilot a viable ChatGPT rival inside M365 guardrails, perfect for proving ROI before your boss questions the spend. Log into the admin center today, enable Anthropic previews, and test Model Council on your next email thread against ChatGPT – capture the before-after to rally your team.
Pulse published by Collab365 Spaces. Cite as "Microsoft adds Claude tools to Copilot for sharper outputs", Collab365 Spaces.