PulseAugur
EN
LIVE 11:34:37

Microsoft Copilot agents fail to perform work tasks reliably

A ZDNet journalist tested Microsoft's premium Copilot agents for Microsoft 365 and Windows, finding them to be "confidently bad" at performing everyday work tasks. Despite occasional moments of competence, the AI frequently produced misinformation, hallucinations, and ultimately wasted time. The experience suggests the AI is not yet ready to reliably handle professional responsibilities. AI

IMPACT Highlights current limitations in AI's ability to reliably perform professional tasks, suggesting a need for further development before widespread adoption for critical work.

RANK_REASON Journalist's review of a product's performance.

Read on Mastodon — fosstodon.org →

AI-generated summary · Google Gemini · from 2 sources. How we write summaries →

COVERAGE [2]

  1. Mastodon — fosstodon.org TIER_1 English(EN) · [email protected] ·

    ZDNet: I paid Microsoft’s premium Copilot agents to do my work – they were confidently bad at it. “Over the past few weeks, I’ve been trying to use the AI featu

    ZDNet: I paid Microsoft’s premium Copilot agents to do my work – they were confidently bad at it. “Over the past few weeks, I’ve been trying to use the AI features in Microsoft 365 and Windows for a variety of everyday work tasks. Copilot shows occasional flashes of competence, b…

  2. Mastodon — fosstodon.org TIER_1 English(EN) · [email protected] ·

    I paid Microsoft's premium Copilot agents to do my work - they were confidently bad at it Can you get a Copilot agent to do your work for you? I tried, but the

    I paid Microsoft's premium Copilot agents to do my work - they were confidently bad at it Can you get a Copilot agent to do your work for you? I tried, but the AI wasn't ready to play along. https://www. zdnet.com/article/microsoft-36 5-premium-copilot-agents-hands-on/ # Tech # T…