AWS Quick Adds Conversational Workflow Building and Multi-Agent Orchestration
Amazon Web Services updated AWS Quick on November 20, adding business workflow creation through chat, parallel multi-agent orchestration, a new web application, and more connectors. Workflows can combine agents, skills, tools, structured data, and fixed computational steps, and can run across web, desktop, and mobile platforms. The update also adds prompt guardrails in preview, connector permissions, and approval for write operations, along with support for Japanese, German, Simplified and Traditional Chinese, and Vietnamese.
Why read this
SelectedAI score 77
The update integrates conversational workflows, multi-agent collaboration, and permission governance into a single product, providing teams with a concrete model for reusing automated workflows.
Topics
Body
Amazon Web Services released a broad AWS Quick update on November 20, 2026, adding chat-built business workflows and orchestration that pools output from several specialized agents. A redesigned web app, new connectors, and expanded administrative controls support a push to make AI-assisted processes repeatable across teams.
Conversational Flows Standardize Recurring Work
Users can describe a process in chat and have Quick construct a flow that they can run, test, and refine in the same conversation. A step diagram beside the chat supports review, and flows can coordinate agents, skills, tools, structured data, knowledge, and applications. They can be started from the web, desktop, or mobile. New deterministic steps handle fixed calculations and checks alongside agent-led drafting and analysis; AWS says these make repeated workflows consistent and verifiable, with support for longer-running processes and larger data volumes.
AWS’s explicitly fictional finance example shows how those pieces fit together. A monthly reporting flow analyzes regions in parallel, compares results with plan using fixed spreadsheet formulas, applies manager-set thresholds to flag significant variances, and drafts commentary. It then prepares an Excel summary and email for review. A person examines why regions were flagged, requests changes in chat, and approves the report before distribution. Colleagues can subsequently run the same flow, reducing dependence on the original employee’s private checklist.
Pomelo, a Latin American fintech provider, supplied a customer example. Head of Data Rodrigo Marini said nearly four times as many people now work with data each month without expanding the data team, one use case saves 80 hours of work per month, and employees across the business can build agents and flows under shared governance. The result illustrates the release’s central business case: broader participation in data work without matching growth in specialist headcount.
Web App, Agents, and Connectors
The updated browser home screen combines a prioritized feed of email and messages with a daily briefing. Users can ask about an item or have Quick draft a reply for review. Scheduled and longer tasks continue after the browser closes, while Mission Control surfaces completed results and agents awaiting input when users return.
Library collects recent work, favorites, pinned items, and Quick-created files, while conversations, agents, and files synchronize across web, desktop, and mobile. A Customize section brings agents, flows, skills, knowledge, and connectors together, including desktop-built and shared agents and skills. Existing dashboards, topics, and spaces continue working without activation steps.
Multi-agent orchestration lets the agent receiving a request call several specialized agents simultaneously. Each retains its own instructions, knowledge, and context, so domain experts can maintain their own agents; Quick then presents a combined response showing each contribution. In AWS’s shipping-cost example, a forecasting agent using planning data and a procurement agent using vendor contracts address different parts of the same question, leaving the user to decide the next action.
The Meta ads connector uses Meta’s ads Model Context Protocol server for natural-language campaign creation, management, and analysis. Amplitude’s MCP server supports questions about product usage, customer behavior, and conversion. Across connectors, existing access permissions still apply, administrators decide which integrations teams can add, and Quick requires permission for write actions such as creating a Meta campaign or changing a budget.
Governance Controls Shape the Rollout
Reusable skills, previously managed on desktop, can now be browsed, installed, created, and managed in the web app. These instructions capture practices such as checking a contract against a team checklist or formatting a project update.
Prompt guardrails are in preview. Administrators can filter content, block prohibited topics and specified words, detect prompt attacks, and protect selected personal-information types. They can test a policy in detect-only mode before enforcing it for individual users, identity-provider groups, or the entire account.
Quick adds Japanese, German, Simplified and Traditional Chinese, and Vietnamese. AWS says the features are available immediately in every Region where Quick is offered, and that customer data remains in the customer’s environment and is not used to train anyone else’s models.
Wider automation therefore comes with explicit operating boundaries: group permissions, connector restrictions, write approvals, and human review define what agents and flows may do when organizations extend Quick across teams.
AI Trends