Amazon unveils a Copilot for all your apps
The Register
Amazon just launched a desktop AI assistant that watches everything you do across email, Slack, and calendars to auto-book meetings for you. It's Amazon's Copilot-killer, and it comes with a side of 'trust us' on privacy.
Based on reporting by The Register — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Amazon spent Tuesday in San Francisco trying to convince the world it can out-Copilot Microsoft, out-agent Salesforce, and out-automate everyone else building AI into the workplace. The centerpiece is a new version of Amazon Quick, a desktop app that Amazon says "lives on your computer and connects directly to your work." No AWS account required, just an email address and a willingness to let the thing watch what you're doing all day.
The pitch is seductive in the way these pitches always are. Amazon Quick VP Jigar Thakkar demoed a scenario where instead of pinging five people across Slack, email, and Zoom to schedule a meeting, you type one line into Quick — set up a meeting for project x — and it handles the rest. It knows who's busy, who's free, and fires off the invite. That's possible because Quick logs your activity constantly, building a running profile of deadlines, collaborators, and whatever else it decides is relevant context. Native hooks into Google Workspace, Microsoft 365, Zoom, and Salesforce make that surveillance frictionless.
When a reporter asked Thakkar about the obvious security implications of a model that never stops watching, he didn't offer specifics. He pointed to AWS's two decades running cloud infrastructure and left it there. Given how much personal and professional data Quick is designed to hoover up, a vague appeal to institutional trust feels thin, though Amazon will presumably publish more detailed documentation eventually.
The second announcement is a much bigger structural bet: Amazon Connect, the contact-center platform Amazon has run since 2017, is splitting into four agentic products. Connect Decisions targets supply chains, Connect Talent handles hiring — Amazon showed off an AI chatbot conducting an actual job interview, which lands somewhere between impressive and unsettling — Connect Health covers healthcare workflows, and Connect Customer AI is simply the old customer-service product with a new name. Each comes packaged with AI "teammates" meant to run business processes semi-autonomously, though Amazon is careful to insist humans stay in the driver's seat.
None of this happens in a vacuum. Salesforce is already selling an agents-everywhere vision, albeit while patching a vulnerability that let Agentforce leak sales data. Google has Gemini Enterprise. Microsoft just rolled out a new 365 tier built around Agent 365, a control plane specifically for managing AI agents. Amazon is arriving to a market that's already crowded with entrenched vendors, and switching costs in enterprise software are notoriously brutal. Building a slicker assistant is the easy part; getting companies to rip out tools they've spent years embedding themselves in is the actual fight.
My take — AI-written commentary, not fact-checked reporting
Every one of these companies wants an AI that watches your entire digital life and calls it "context," and every one of them answers security questions with vibes instead of specifics — Amazon's own VP literally pointed to AWS's resume instead of explaining anything. I don't think the agent-everything land grab is really about productivity; it's about who owns the plumbing of your workday, and locking you into their surveillance layer is the actual product.
Read more about this at: The Register