Microsoft Unveils "Scout," an AI Coworker That Never Logs Off, and Project Solara for Agent-First Devices — AI News Roundup for June 3

Microsoft Unveils AI Assistant "Scout" — Running in Teams as a "Coworker That Never Logs Off"

At its Build conference, Microsoft announced a new AI assistant called Scout. Inspired by the philosophy behind OpenClaw, the assistant is integrated into the Microsoft 365 system and will appear in Teams just like a human colleague.

According to Wired, Scout functions as an "AI coworker that never logs off," automating everyday office tasks. Where the traditional Copilot was visible only to the user, Scout behaves as a visible presence across the organization — a major difference.

Bloomberg has also covered it as an "always-on executive assistant," positioning it as the next evolution of the chatbot. Meanwhile, 404media obtained internal documents and reported that Microsoft aims to make Scout something "users become dependent on." The convenience of AI agents, and concerns about the dependency they create, are likely to spark fresh debate.

According to TechCrunch, Scout operates across the entire Microsoft 365 environment, autonomously managing schedules, gathering information, and executing tasks.

Microsoft's "Project Solara" — A Device Platform for Running AI Agents Natively

Also announced at Build, Project Solara is a platform for running AI agents natively on devices, envisioning a device experience centered on "agents" rather than traditional "apps."

According to Bloomberg, Project Solara aims to take AI agents beyond the laptop screen and embed them in enterprise mobile devices. It has even been described as "putting AI on employees' badges," making it an ambitious project to fuse physical devices with AI agents.

GeekWire analyzes that the platform will bring about a paradigm shift in which "AI agents run in place of apps."

Microsoft's "Majorana 2" — A Next-Generation Quantum Chip with Development Accelerated by Agentic AI

Microsoft has announced its next-generation quantum chip, Majorana 2. The company leverages agentic AI from Microsoft Discovery in the quantum device manufacturing process, reportedly contributing significantly to more efficient materials research.

According to Bloomberg, Microsoft is targeting practical deployment "in 2029" with this technology. Because key components of quantum devices are designed at the atomic level, they traditionally required enormous trial and error, but agentic AI simulations can reportedly cut the number of experiments dramatically.

NVIDIA × Microsoft — Joint Announcement of an Integrated Stack for Agentic AI

NVIDIA and Microsoft have jointly announced an integrated stack for deploying agentic AI. The full-stack solution combines fast hardware, a secure runtime, a responsive data layer, and models optimized for long-running inference, with deployment envisioned across Windows devices, local environments, and the cloud.

In a blog post, NVIDIA emphasized that realizing agentic AI takes "more than a good model." The deep collaboration between the two companies was highlighted as a comprehensive approach spanning hardware to software.

OpenAI Adds Role-Based Plugins to Codex — Toward a General-Purpose App for Non-Developers

OpenAI has expanded Codex's capabilities, adding role-based plugins for data analysis, sales, and more. According to The Decoder, this is part of an effort to make Codex usable as a general-purpose app even by non-developers.

The move aims to extend Codex from a programming tool into a general-purpose business app, making clear a strategy to widen the audience for AI tools beyond engineers.

Google Rolls Out Detection for Deepfake Voice Scams

Google has begun offering a "fake call detection" feature that protects users from impersonation scams using AI deepfakes.

According to TechCrunch, as smartphone users stop answering calls from unknown numbers, scammers are shifting to spoofing trusted numbers and using AI deepfakes to mimic the voices of authority figures or family members. Google's new feature is designed to counter these threats.

As the misuse of AI-generated technology becomes a growing social problem, platforms are accelerating the rollout of protective technologies.