​
What is MCP?

Model Context Protocol (MCP) is an open standard that enables AI assistants to securely connect to external tools, data sources, and services. In IllumiChat, MCP integrations allow your AI assistants to access real-time information from your existing systems and workflows.

​
How MCP Works

1

Assistant receives user query

A user asks a question that requires external data or action.
2

Assistant identifies need for MCP

The assistant determines that it needs to access external information or perform an action.
3

MCP connection established

The assistant securely connects to the configured MCP server.
4

Data retrieved or action performed

The MCP server fetches the required data or executes the requested action.
5

Response generated

The assistant incorporates the external data into its response to the user.

​
Benefits of MCP Integrations

​
Real-Time Data Access

  • Live Information: Access current data from your systems
  • Dynamic Responses: Provide up-to-date information to users
  • Contextual Answers: Give responses based on actual system state

​
Enhanced Capabilities

  • Action Execution: Perform tasks in external systems
  • Data Retrieval: Fetch specific information on demand
  • Workflow Integration: Connect to existing business processes

​
Security and Control

  • Secure Connections: Encrypted communication between systems
  • Access Control: Granular permissions for different operations
  • Audit Trails: Track all MCP interactions and data access

​
Supported MCP Integrations

​
HubSpot Integration

Connect your assistant to HubSpot CRM and marketing data:

HubSpot MCP

  • Access contact and company information
  • Retrieve deal and pipeline data
  • Create and update records
  • Send emails and track engagement

​
Jira Integration

Integrate with Jira project management:

Jira MCP

  • Retrieve issue and project information
  • Create and update tickets
  • Track project progress
  • Manage workflows and transitions

​
Custom MCP Servers

Build your own MCP integrations:

Custom MCP

  • Connect to internal systems
  • Access proprietary data sources
  • Integrate with custom APIs
  • Build specialized workflows

​
Setting Up MCP Integrations

​
Prerequisites

Before setting up MCP integrations, ensure you have:
  • An active IllumiChat account
  • Access to the external system you want to integrate
  • Appropriate permissions and API credentials
  • Understanding of the data you want to access

​
Configuration Process

​
Method 1: Connect via Assistant Configuration

1

Access MCP Settings

Navigate to your assistant’s edit page and click on the “Tools” tab to find the MCP integration section.Assistant Provider List
Important: You cannot connect MCP tools to the default “IllumiChat” assistant. Make sure you’re configuring a custom assistant.
2

Initiate Connection

Click “Connect” next to the desired service (e.g., Jira). This will open the OAuth authorization modal.Jira OAuth Modal
3

Authorize Access

Select your account from the dropdown menu and click “Accept” to grant the necessary permissions.Jira OAuth Accept
4

Confirm Success

Verify that the connection was established successfully. You should see a confirmation message.Jira Connected Success
5

Configure Permissions

Set up what data and actions your assistant can access in the connected system.

​
Method 2: Connect via Workspace

You can also connect MCP tools directly from the workspace:
  1. Open your workspace
  2. Click the ”+” button
  3. Select “Connect MCP Tools”
  4. Follow the same connection process as above
Connect via Workspace
Connecting via workspace is convenient when you’re already working with your data and want to add new integrations on the fly. Just make sure you’re on the correct assistant - you cannot connect MCP tools to the default IllumiChat assistant.

​
Security Considerations

​
Authentication

  • API Keys: Secure storage and rotation of credentials
  • OAuth Flows: Use OAuth 2.0 for enhanced security
  • Token Management: Automatic token refresh and validation

​
Data Privacy

  • Selective Access: Only request necessary data
  • Data Minimization: Limit data retention and usage
  • User Consent: Respect user privacy preferences

​
Access Control

  • Role-Based Permissions: Different access levels for different users
  • Audit Logging: Track all data access and modifications
  • Rate Limiting: Prevent abuse and ensure system stability

​
Best Practices

​
1. Start Small

Begin with simple data retrieval before adding complex actions and workflows.

​
2. Test Thoroughly

  • Test with various user scenarios
  • Verify data accuracy and timeliness
  • Check error handling and fallbacks
  • Monitor performance impact

​
3. Document Integration

  • Document what data is accessed
  • Explain how the integration works
  • Provide troubleshooting guidance
  • Keep documentation updated

​
4. Monitor Usage

  • Track integration performance
  • Monitor error rates and response times
  • Analyze usage patterns
  • Optimize based on insights

​
Common Use Cases

​
Customer Support

  • Access customer information from CRM
  • Retrieve order and purchase history
  • Create support tickets automatically
  • Update customer records

​
Sales Enablement

  • Get real-time deal information
  • Access product catalogs and pricing
  • Create leads and opportunities
  • Track sales activities

​
Project Management

  • Retrieve project status and progress
  • Create and update tasks
  • Access team information
  • Track time and resources

​
Data Analysis

  • Access business intelligence data
  • Generate reports on demand
  • Provide real-time analytics
  • Answer data-driven questions

​
Troubleshooting

​
Common Issues

​
Next Steps

Data Retrieval Methods

Learn about the differences between knowledge base and MCP data access.

Workspace

Use the workspace for private data analysis with MCP tools.

Assistant Configuration

Configure your assistant’s MCP integrations and tools.

Creating Assistants

Learn how to create assistants that can use MCP tools.