How to Connect Enterprise Systems to Amazon Quick Desktop with MCP
Ismael Murillo, Solutions Architect — Generative AI, Amazon Web Services
Use Case and Problem
Your team’s data lives in multiple enterprise systems CRM platforms, ticketing tools, HR databases, knowledge bases, and more. Answering a simple question like “Which accounts are up for renewal next month?” often means logging into a separate system, running a report, and copying the results. What if you could query any connected system from the same chat window where you read your email and manage your calendar?
Amazon Quick Desktop supports MCP (Model Context Protocol) an open standard that lets AI assistants connect to external data sources and tools. With MCP, you can query databases, call APIs, and pull data from enterprise systems directly from the Amazon Quick chat. This article walks you through connecting an MCP server and running your first cross-system queries.
How to Solve the Problem
This article shows you how to connect an MCP-compatible data source to Amazon Quick Desktop and query enterprise data using natural language.
Prerequisites
· Amazon Quick Desktop installed and signed in
· An MCP server endpoint (self-hosted or provided by your organization)
· Network access to the MCP server from your machine
What is MCP?
MCP (Model Context Protocol) is an open standard that defines how AI assistants communicate with external tools and data sources. Think of it as a universal adapter — any system that exposes an MCP server can be queried by Amazon Quick Desktop without custom integration code.
Common MCP server examples:
- Database connectors — query PostgreSQL, MySQL, DynamoDB, or Redshift
- SaaS integrations — pull data from Salesforce, Jira, ServiceNow, Zendesk
- Knowledge bases — search internal documentation stored in S3 or other repositories
- Custom APIs — any REST API wrapped in an MCP server
Step 1: Add an MCP Connection
1. Open Amazon Quick Desktop → Settings → Capabilities → Connections
2. Scroll to the MCP Servers section
3. Click “Add MCP Server”
4. Choose your Connection type:
5. Local — Runs a command on your machine (ideal for locally-hosted MCP servers)
6. Remote — Connects to a remote endpoint via URL
7. Fill in the connection details:
8. ID: A unique identifier for the server (e.g., aws-sentral-mcp)
9. Name: A friendly display name (e.g., “AWSentral MCP”)
10. Command: The executable to run (e.g., aws-sentral-mcp)
11. Arguments: Command line arguments (e.g., -m mcp_server --port 8080)
12. Description: What this connection provides (e.g., “Salesforce CRUD for accounts, opps, and SA activities”)
13. Click Connect and wait for the connection to register
Tip: Your IT or data engineering team can set up MCP servers for internal databases. The MCP specification provides server templates for common data sources. For local connections, the MCP server binary must be installed on your machine.
Step 2: Verify the Connection
Once connected, verify Amazon Quick can access the MCP server by asking a simple question:
What data sources are available?
Amazon Quick will list all connected data sources organized by category — Local Folders, Connected Services, Cloud/Enterprise Tools, Knowledge & Search, and Specialized Tools — confirming the integration is working.
Step 3: Query Enterprise Data with Natural Language
Now you can query your enterprise systems directly from the chat. Here are practical examples:
Example 1 — Customer account lookup:
Look up Netflix in AWSentral. Show me the account details.
Amazon Quick queries the connected CRM through MCP and returns structured account data in a rich table format — complete with metadata badges showing the source system, response time, and result count. No need to log into Salesforce separately.
Example 2 — Cross-system analysis:
Show me all open support tickets for accounts that have
a renewal coming up in the next 60 days. Flag any with
a satisfaction score below 3.
This query crosses two data sources — the support ticketing system and the CRM — and Amazon Quick joins the results automatically.
Example 3 — Knowledge base search:
Search our internal knowledge base for the latest
guidelines on data migration best practices.
Amazon Quick queries an S3-backed or vector database knowledge base through MCP and returns relevant documentation.
Step 4: Build Artifacts from Enterprise Data
The real power comes from combining MCP data with Amazon Quick’s builder capabilities:
Using my opportunity data from AWSentral, build me a
portfolio dashboard showing total opportunities, launched
revenue, stage distribution, and primary competitors.
Amazon Quick pulls live data through MCP, analyzes it, and generates an interactive dashboard — all from a single prompt. The resulting artifact shows real-time KPIs, charts, and tables sourced entirely from your connected enterprise systems.
You can also create documents:
Draft a customer health report for our top 10 accounts
by revenue. Include their NPS scores, open support tickets,
and renewal dates. Format as a Word document.
Step 5: Combine MCP with Agents
For maximum automation, create an agent that leverages MCP connections on a schedule:
Create an agent that runs every Monday at 8 AM and:
- Pulls all accounts with renewals in the next 30 days from our CRM
- Checks the support system for any open critical tickets on those accounts
- Generates a renewal risk summary with recommended actions
This creates a recurring workflow that queries enterprise systems automatically and delivers insights without manual effort.
**
Troubleshooting
**
| Issue | Solution 2 | ||
|---|---|---|---|
| Connection fails | Verify the MCP server binary is installed and the command is in your PATH. Check VPN if connecting to internal systems. | ||
| Authentication error | Confirm credentials are current. API keys may expire — check with your admin. | ||
| Slow responses | Large queries may take time. Try narrowing your request with filters (date ranges, specific accounts). | ||
| No data found | Verify the MCP server has the expected schema. Ask: What tools are available in [connection name]? | ||
| Local server won’t start | Check that the Command and Arguments fields are correct. Try running the command manually in your terminal first. |
Conclusion
MCP connections extend Amazon Quick Desktop from a personal productivity tool into an enterprise data hub. By connecting your CRM, ticketing systems, knowledge bases, and databases through MCP, you can query any system using natural language — without switching applications or writing code. Start with one connection, prove the value, and expand from there. The combination of MCP data access, interactive dashboard generation, and scheduled agents creates workflows that would otherwise require multiple tools and manual effort.
For more on MCP and available server implementations, visit the Model Context Protocol documentation.
Author Bio
Ismael Murillo is a Worldwide Generative AI Solutions Architect at Amazon Web Services, specializing in Amazon Quick. Based in Tempe, AZ, he is a combat military veteran, cancer fighter, and passionate advocate for making AI tools accessible to every knowledge worker. He leads QuickSight Community User Groups across the US.
| Issue | Solution 2 | ||
|---|---|---|---|
| Issue | Solution | ||
| Connection fails | Verify the MCP server binary is installed and the command is in your PATH. Check VPN if connecting to internal systems. | ||
| Authentication error | Confirm credentials are current. API keys may expire — check with your admin. | ||
| Slow responses | Large queries may take time. Try narrowing your request with filters (date ranges, specific accounts). | ||
| No data found | Verify the MCP server has the expected schema. Ask: What tools are available in [connection name]? | ||
| Local server won’t start | Check that the Command and Arguments fields are correct. Try running the command manually in your terminal first. |
Suggested tags: amazon-quick, mcp, integrations, howto, enterprise, data-connectivity