AWS launches Amazon Quick, a desktop AI agent for background work and office tasks

AWS has unveiled Amazon Quick, a desktop AI agent that runs directly on the user's computer, is not limited to a single chat window, and gradually accumulates a personal work context. The idea is that the assistant does more than answer questions: it pulls data from files, email, calendars and corporate services on its own, and turns long office processes into tasks that take a few minutes.

How Quick works

Amazon's main bet is persistent context. Quick lives on the laptop, has access to local files, can be linked to a calendar, email and work apps, and then uses those signals to better understand the user's current tasks. The company says the assistant learns from every session, builds a personal knowledge graph, and over time adapts more precisely to the user's working style, circle of contacts and business context.

Amazon stresses that Quick does not lock itself into its own ecosystem. It can work with several widely used services at once, including Google Workspace, Microsoft 365, Slack, Zoom, Salesforce, Outlook, Gmail, Teams, Jira and ServiceNow. The agent can also automate browser workflows and connect to developer tools such as Kiro CLI and Claude Code, so it can not only analyze information but also run chains of actions across different environments.

What the agent can do

Amazon positions Quick as a personal work companion for sales, marketing, finance, operations and development. In effect, the company is trying to combine chat, data search, BI tools, content generation and routine automation in a single interface. Thanks to persistent context, the assistant is supposed to do more than answer requests: it should understand what the team needs at a given moment and which sources it can rely on, without manually copying files and links.

• Write emails, prepare follow-up reminders and fill in the right context from previous correspondence
• Create documents, presentations, infographics and images directly from the chat
• Surface relevant Slack threads, notes and files before meetings without a separate request
• Build dashboards and analytical conclusions from live corporate data
• Automate repetitive actions in the browser and inside work processes

Amazon has also made onboarding easier: new users can sign up in a few minutes with a personal email or through existing Google, Apple, GitHub or Amazon accounts, without a separate AWS account. The lineup includes free and paid plans, and companies can choose Professional and Enterprise tiers with broader agent and BI capabilities. For mass adoption this matters: Amazon is clearly trying to lower the barrier to entry and get Quick into teams' daily work faster.

A bet on the enterprise

Quick is being promoted not as yet another general-purpose chatbot but as a corporate layer on top of existing infrastructure. Amazon emphasizes that the service connects to company data while preserving the access permissions already configured, runs on AWS infrastructure, and supports the same encryption, IAM and compliance mechanisms that enterprises already use today. The company also promises that customer data is not used to train other parties' models.

Teams get shared Spaces, where they can share dashboards, agents, automations and accumulated knowledge. The idea is simple: if one employee builds a useful workflow or dashboard, the whole team can use it, not only its author.

According to Amazon, the product is already being tested by 3M, GoDaddy, AstraZeneca, BMW, New York Life, Southwest Airlines and Amazon itself. Inside Amazon Books, the time needed to prepare coordination documents fell by 80%, and engineering teams cut factory testing time by 67%.

"Quick helps our people complete tasks in minutes, not hours," is how

Mondelez described the effect.

What this means

Amazon is entering the race not with just another chat assistant but with an attempt to become a permanent AI layer on top of the desktop and corporate applications. If Quick can reliably hold context, work with data safely and consistently automate actions across services, it will be more than a "text helper": it will be a new interface for everyday office work.