Amazon Quick Plus vs Professional: Lessons From a Month of Real Usage
What happens when Amazon Quick becomes part of your daily workflow and you hit the wall? Lessons from a month of real usage, hidden plan differences, MCP-to-Lambda patterns, and the surprisingly fast origin story behind Amazon Quick.
Amazon Quick Plus vs Professional: Lessons From a Month of Real Usage
I signed up for Amazon Quick Plus to build custom MCP integrations and use Deep Research as a daily work tool. $25/month, no AWS account required — seemed like a good deal.
After 14 days, I hit 93% of my agent hours quota. Quick sent a friendly warning. A few days later: no more AI features until the next billing cycle. No option to buy additional hours. No top-up. Just wait.
That's when I discovered the real differences between the plans. Not the ones on the feature page — the ones you only notice when you hit a wall. The $250 infrastructure fee for Professional isn't just about governance and SSO. It's also the only path to buying more agent hours when you run out. And that distinction doesn't jump out from AWS's plan comparison.
Here's the practical guide I wish I'd had on day one.
Two Sign-Up Paths, Four Plans
The most important distinction in Amazon Quick isn't a feature — it's how you sign up.
Path 1: quick.aws.com — Sign up with email, Google, Apple, GitHub, or Amazon credentials. No AWS account required. No infrastructure fee.
Path 2: AWS Management Console — Deployed into your AWS account. $250/month infrastructure fee. Full governance, compliance, and enterprise controls.
| Plan | Sign-Up | Price/User/Month | Infrastructure |
|---|---|---|---|
| Free | quick.aws.com | $0 | — |
| Plus | quick.aws.com | $25 ($20 annually) | — |
| Professional | AWS Console | $20 | $250/month |
| Enterprise | AWS Console | $40 | $250/month |
The per-user price of Plus (monthly) is actually higher than Professional — but Professional adds a $250 base fee for governance, QuickSight integration, and AWS Support. Note: I'd recommend paying monthly for Plus rather than committing annually. Given the hard quota limit and no top-up option, you'll want the flexibility to cancel if it doesn't fit your usage pattern.
What's Shared Across All Plans
Every plan — including Free — gives you the core: chat with AI, Deep Research, Spaces, Quick Flows, a no-code app builder, and connectors to Slack, Microsoft 365, Google, and QuickBooks. The basic integrations are not a differentiator. They work everywhere.
What Separates the Plans
The differences come in three layers:
Plus ($25/month) adds team features: custom chat agents, shared Spaces, shared knowledge bases, the desktop app, and browser extensions.
Professional ($20 + $250 infra) adds governance: QuickSight dashboard viewing, SSO, RBAC, data sovereignty, centralized billing, unlimited user scaling, and 24/7 AWS Support. This is the AWS Console path — per-user price actually lower than Plus monthly, but with the $250 infrastructure fee.
Enterprise ($40 + $250 infra) adds creation and automation: build QuickSight dashboards and datasets, Quick Automate with UI Agent, full connector creation, double the agent hours (8h vs. 4h), and double the index storage (50 GB vs. 25 GB per user).
Three Interfaces, One Account, Different Capabilities
Amazon Quick isn't just one app. It ships as three separate interfaces — all tied to the same account, but with different feature sets:
- Web App (quick.aws.com) — the full experience: Deep Research, Spaces, Flows, knowledge bases, conversation history.
- Desktop App (Windows/Mac) — Activity Feed monitoring email and calendar, chat, local file access. Some features like Deep Research are only available through the web app.
- Mobile App (Android/iOS) — a bare-bones chat interface. At the time of testing, I couldn't see previous conversations, pinned chats, or data from the web or desktop apps. It's essentially a standalone chat window with no shared history visible.
Same account, same quota — but not the same experience everywhere. If you're expecting the mobile app to mirror the desktop or web app, you'll be surprised by how stripped-down it currently is.
Agent Hours: The Usage Limit That Matters
Note: As of June 2026, Plus is the most comprehensive plan on quick.aws.com with no option to purchase additional agent hours. This may change — but at the time of writing, the hard limit is real.
All plans work with "agent hours" — compute time consumed by AI features like the Desktop app, custom apps, Research, Flows, and Automate. Usage is metered per second.
For Professional and Enterprise, AWS publishes clear quotas:
- Professional: 4 hours/month (2h agents + 2h research)
- Enterprise: 8 hours/month (4h agents + 4h research)
- Additional hours purchasable: $3 (Flows/Automate) or $6 (Research)
For Plus? The exact quota isn't publicly documented. The pricing page doesn't mention a number. You find out you have a limit when you get the 80% warning notification — and you never learn exactly how many hours you started with. In my case, mostly chat and Desktop app usage (barely any Deep Research) exhausted the quota in about 14 days.
What Happens When You Hit the Limit
Quick warns you when you cross the 80% threshold:
At this point, you start rationing. Shorter prompts, skip Deep Research, think twice before triggering a Flow. Within days, you blow through the remaining 20%. Then it's over:
No error page. No upgrade prompt. Just the chat interface telling you to come back in 12 days. The AI features you've been building around — Research, Flows, custom agents — are simply unavailable until the quota resets.
What Still Works (and What Doesn't)
After hitting the limit, Quick doesn't shut down completely. Here's what I found:
Still works:
- Activity Feed (Desktop app) — continues monitoring email and calendar. You still get summaries and can react to items, just not via chat.
- Exporting conversations — you can export previous chats anytime and continue a topic in another AI. You lose Quick's skills and your knowledge base context, but the content isn't trapped.
- Local files — Quick lets you create and access local markdown files at any time. If they're in a shared folder or "My Stuff," you can simply copy them out.
Credit where it's due: AWS doesn't lock you out of your data when the quota runs dry. You can still export chats and files, pick up the thread elsewhere. The Quick-specific context — knowledge bases, agent configurations, conversation history as context — is gone, but at least the raw content remains portable.
Completely blocked:
- All AI chat — any chat input is locked. No agent tasks, no follow-up questions, no "just one quick prompt."
- Deep Research (web app) — locked, with an "Upgrade Plan" button that leads nowhere useful.
The Research lock is particularly frustrating: the upgrade prompt suggests there's a path forward, but Plus is already the highest plan on quick.aws.com. The only real upgrade is jumping to the AWS Console path at $250+/month.
Can you buy more agent hours? Only from Professional onwards. In Plus, this option simply doesn't exist. Your only recourse: be more frugal next month.
In practice, the upgrade path looks like this:
- Plus ($25/month): Hard limit. Quota exhausted = wait until next billing cycle.
- Professional ($20/month + $250 infra): Top-up at $3/agent hour or $6/research hour.
- Enterprise ($40/month + $250 infra): Double the quota (8h vs. 4h) and top-up available.
The Practical Impact
If you use Quick seriously as a daily work tool — Research, Flows, Desktop app — you can hit the Plus limit in two weeks, not a month. And then you face an awkward set of choices:
- Ration: Use Quick only for the most important tasks. Less context, simpler queries, skip Deep Research.
- Upgrade to Professional/Enterprise: Jump from $25 to $270+ per month. Rarely practical for an individual.
- Export and switch: Move your content, knowledge bases, and workflows to another tool. And hope the remaining quota is enough to do so.
This creates a practical consideration: as you build knowledge bases and workflows inside Quick, they become harder to move when you're out of quota. Plan for this before it happens.
Connectors: Using vs. Creating
At first glance, connectors look the same across plans. The difference is in who can set them up and how:
- Free/Plus: Use the pre-built integrations (Slack, M365, Google, QuickBooks). Build custom MCP servers via the Coding Agent Skill — but only source code; hosting is on you.
- Professional/Enterprise: Use connectors shared with you. Register and deploy custom connectors through Quick — managed hosting, credential management, and team sharing included.
MCP and the Coding Agent Skill
Amazon promotes MCP (Model Context Protocol) as an open standard — "you're never locked into a walled garden." Initially, I assumed this was plan-gated. After digging deeper, the picture revolves around Quick's Coding Agent Skill — and the distinction between using MCP servers and building your own.
Using MCP Servers
All plans can use existing MCP servers. Quick connects to public and internal MCP servers — Bedrock supports an MCP Gateway for curated, monitored access. If there's already an MCP server for your use case (Atlassian, Asana, Monday.com, or an internal one your company runs), you can connect it.
Building Your Own MCP Servers
This is where the Coding Agent Skill comes in. Quick Desktop includes this skill, which connects via ACP (Agent Client Protocol) to a local client like Kiro CLI. It can build MCP servers for you — for example, to run defined queries against your data sources or wrap an API you use daily.
The Plan Difference
In Free and Plus: You can use existing MCP servers, and you can use the Coding Agent Skill to build new ones. But Quick only generates the source code — hosting, credentials management, deployment, and scaling are entirely your responsibility.
In Professional and Enterprise: Same building capabilities, but Quick handles deployment and hosting for you. MCP servers you build through the Coding Agent Skill are directly hosted, versioned, and shareable within your team. Same for applications — build, host, share, connect to approved data sources. This is the old QuickSight heritage showing through: dashboards and data apps, but now AI-generated.
The Practical Trade-Off
On Free/Plus, you can build a custom Google Tasks MCP server through the Coding Agent Skill — but you're responsible for running it yourself (locally or on EC2/ECS). On Professional/Enterprise, Quick deploys, hosts, and shares it for you, with credential management and team access built in. Classic AWS pattern: build it yourself on any plan, pay for the managed experience.
The takeaway: MCP isn't locked to higher plans — neither using nor building. What's plan-gated is the lifecycle: deployment, hosting, sharing, versioning, and credential management. If you're comfortable running your own infrastructure, Free/Plus gets you further than the feature page suggests.
When MCP Gets Expensive: The Graduation Pattern
Here's something I picked up from an AWS team talk: MCP is great for prototyping automations, but every MCP call runs through an LLM — which means it consumes agent hours and model costs on every invocation.
Their real-world example: the Activity Feed notices a recurring pattern — say, parsing team members from an email and generating calendar invites for everyone. First iteration: an MCP server handles it, the agent interprets the email, extracts names, and orchestrates the invites. Works great. But when this happens 20 times a day, the LLM costs add up fast.
The solution: graduate the automation. Use the Coding Agent Skill to write a Lambda or Step Function that does the same thing deterministically — parse the email, extract the team, send invites — without an LLM in the loop. Same result, fraction of the cost.
The pattern looks like this:
- Discover — Activity Feed surfaces a recurring task
- Prototype with MCP — Quick orchestrates it via agent + MCP server
- Validate — it works, you use it daily, agent hours burn fast
- Graduate — Quick's Coding Agent builds a Lambda that replaces the MCP call with direct API logic
All orchestrated inside Quick. The AI builds its own replacement for the expensive parts. This also explains why agent hours vanish so quickly on Plus — if you're running MCP-heavy workflows daily, you're burning quota on every single invocation.
Quick Flows vs. Quick Automate
The resource hub articles on builder.aws clarify a distinction that's easy to miss:
Quick Flows (all plans):
- Prompt-generated multi-step automations
- Steps include: Chat Agent, Research, Web Search, UI Agent, Create Image, Data Insights, Input Forms
- Schedulable, shareable with teams
- Use case: weekly reports, ad copy generation, data analysis pipelines
Quick Automate (Enterprise only):
- Multi-agent automation for complex enterprise processes
- UI Agent for browser automation (navigates websites dynamically via natural language)
- Credential store for secure secret management
- Human-in-the-loop approval gates
- Cross-department, cross-system workflows
- Use case: legal response automation, logistics control towers, Salesforce pipelines
The naming is similar, but the capability gap is significant. Quick Flows is your daily task automation — available everywhere. Quick Automate is deployed, multi-agent enterprise automation with governance built in.
Why It's Confusing: Two Products Became One Name
If you've used AWS for a while, you might remember QuickSight (managed BI — dashboards, datasets, data visualization) and Amazon Q Business (AI assistant for enterprise knowledge work). These were separate products with separate audiences: QuickSight served data teams, Q Business served knowledge workers.
At re:Invent 2025, AWS merged them into QuickSuite. In Q2 2026, it became Amazon Quick — with Free and Plus plans opening up on quick.aws.com.
The Desktop app itself has a remarkable origin story. According to GeekWire , in late January 2026, AWS VP Swami Sivasubramanian pulled together a team of about six engineers after realizing the underlying models had gotten good enough — the missing piece was connecting them to the systems where people actually work (email, calendar, Slack, documents). Six weeks later, 200 people inside Amazon were using it. Ten weeks in, 10,000. They shipped publicly on April 28 — three months after they got started. The classic Amazon PRFAQ was written after the product was already in beta.
Here's why this matters for you: the two sign-up paths (quick.aws.com vs. AWS Console) aren't a design choice. They're a legacy of two fundamentally different products being unified under one brand. The quick.aws.com path descends from Q Business — chat, research, connectors, no AWS account needed. The AWS Console path descends from QuickSight — dashboards, datasets, governance, IAM integration.
That's why the plan structure feels strange. It's not one product with four tiers. It's two products wearing the same name, built by small teams moving fast. The feature page presents them as a single offering — but the moment you look at deployment, billing, and capabilities, the seam is visible.
Decision Guide
| Your situation | Recommended plan |
|---|---|
| Individual exploring AI productivity | Free |
| Individual or small team, daily use | Plus (watch the quota) |
| Team needing governance, SSO, QuickSight | Professional |
| Enterprise with compliance requirements, heavy automation | Enterprise |
| Just want custom MCP integrations | Free/Plus (build & self-host) or Professional/Enterprise (managed deployment, hosting, sharing) |
| Need to buy additional agent hours | Professional minimum |
The Rules of Thumb
You want Quick as a personal AI assistant? → quick.aws.com, Free or Plus. Done.
Your team needs compliance, SSO, and QuickSight? → AWS Console, Professional or Enterprise. The $250 is the price of governance.
You're solo or a small team with an AWS account? → Still go with quick.aws.com. The AWS Console path only pays off when you need governance features or the ability to buy more agent hours.
You're hitting the Plus quota regularly? → Consider whether the $250 jump to Professional is justified by the ability to top up — or whether rationing and supplementing with other tools is the more practical path.
Sources: Amazon Quick Pricing , Amazon Quick Features , AWS Docs: Plans , Amazon Quick Resource Hub , GeekWire: How agentic AI is rewiring Amazon's teams — accessed June 2026
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article