Copilot Settings
The latest Copilot Settings coverage — news, analysis, and updates from the WindowsNews.AI desk.
Mistaken 'Thank You' Email Thwarted Apple-OpenAI Trade Secret Talks Before Lawsuit
Apple's trade-secret lawsuit against OpenAI takes an embarrassing turn as emails show an outside attorney's misdirected thank-you message derailed settlement talks before they started, contradicting Apple's claim that OpenAI never responded to its concerns. The case has no direct impact on Windows or Microsoft services, but it highlights the messy realities of tech legal disputes.
Delivery Giant Evri Puts Microsoft’s New AI Agent Bundle to the Test with 6,000 Users
UK parcel carrier Evri is deploying 6,000 Microsoft 365 E7 licenses, putting Microsoft Copilot, Agent 365, and the Entra Suite into workers' hands. The early large-scale rollout builds on Evri's Power Platform automation wins and tests whether governed AI agents can deliver real operational gains in logistics while keeping data secure.
Microsoft 365 Copilot Could Get Detail Controls After Neurodivergent Employees Report Overwhelm
Microsoft has launched a formal product-feedback program that puts neurodivergent employees directly into the design and testing loop. Early sessions found that Microsoft 365 Copilot can overwhelm users with too much information, prompting teams to explore detail controls, smarter defaults, and adaptive personalization. While no features have shipped, the initiative signals a shift toward cognitive ease as a design priority for AI interfaces.
AI-Powered Training Lands at Axalta: 4,000 Courses, Azure AI, and Entra ID in 90 Days
Coatings giant Axalta deployed a global, AI-driven training platform in just 90 days using Unboxed Technology and Microsoft Azure. The system delivers 4,000 courses with AI role-play and coaching, tied together with Entra ID for identity. While the vendor reports large skill gains, IT leaders must scrutinize data governance and metrics before following suit.
Microsoft's AI Agent Security Playbook: Give Them an Identity, Not a Master Key
Microsoft's July 16 security guidance urges organizations to abandon broad permissions for AI agents and instead enforce unique identities, task-scoped RBAC, tool allowlists, and just-in-time elevation. The playbook addresses real-world incidents where agents gained excessive access, and lays out a 30–90 day action plan for IT admins, developers, and business leaders.
Intel Arms Engineers with Google's Gemini AI to Speed Next-Gen Chip Design
Intel announced an internal deployment of Google Cloud’s Gemini Enterprise AI and high-performance computing to speed chip design and automate workflows. While no consumer-facing changes are coming, the move could shorten development cycles for future Intel processors. IT admins and developers have no immediate action to take, but the partnership signals a growing reliance on cloud and AI in semiconductor manufacturing.
Acer's Swift Go 14 AI Tested: 21-Hour Battery in a Sub-$1,300 OLED Laptop
The Acer Swift Go 14 AI (2026) achieves 20 hours and 51 minutes of battery life in testing, thanks to Intel's efficient Panther Lake platform. Priced at $1,299, it combines a vibrant 14-inch OLED display, a best-in-class port selection, and lightweight design, though the 60Hz screen and limited storage may give some buyers pause.
Microsoft’s AI-Driven Support Content Manager Could Save 16,000 Hours a Year—and It’s Coming to Your Team
Microsoft's internal IT organization has built an AI pipeline that scans support tickets to automatically identify gaps in knowledge bases and draft new articles, projecting 16,000 saved hours annually. The system, still in internal testing, will be released to customers, offering IT teams a blueprint for automating their own knowledge management.
Microsoft Lays Out AI Railroad Brain to Solve Freight Rail’s Coordination Crisis
Microsoft has proposed an AI Railroad Brain concept that uses digital twins, real-time data, and AI copilots to coordinate freight rail decisions across dispatching, maintenance, safety, and fuel efficiency. The plan targets five chronic rail challenges and advises operators to start with a narrow use case, connect relevant data, and embed AI into existing workflows with clear human oversight.
Stop the Silent AI Data Leak: 5 Questions Your Vendor Must Answer Before Purchase
A Spiceworks report warns that enterprises are adopting AI tools faster than their governance can keep up, leading to accidental data exposure through shadow AI. IT teams must demand concrete answers on data training, residency, and retention before purchase, or risk proprietary information becoming public training data.
Australia to Data Centres: Fund Your Own Power or Don't Build — What the 2027 Rules Mean for Cloud
Australia's proposed data-centre regulations require operators to self-fund new power generation and grid connections, potentially reshaping cloud expansion. The July 2026 framework triggered moratorium calls from environmental groups, but the government aims for legislation by early 2027. Here's what the changes mean for businesses, cloud users, and the future of AI infrastructure in Australia.
Canva Code 2.0 Unlocks HTML Import and Drag-and-Drop Editing for Free Users
Canva has opened its Code 2.0 tool to free and education users, enabling anyone to build interactive websites, apps, and games from plain-language prompts. The update adds HTML import, visual editing, and faster performance, though freeform drag-and-drop remains absent.
Patch in 72 Hours: Microsoft's New Windows 11 Rule Is a Reaction to AI-Powered Attacks
Microsoft now urges enterprises to deploy Windows 11 quality updates within three days of release, a major shift from traditional weeks-long testing cycles. The recommendation follows record-sized Patch Tuesday arrivals and AI-driven acceleration of vulnerability discovery and exploitation. IT teams must modernize with automated deployment rings, while consumers should simply keep automatic updates on.