Integration guide

Connect CRM and ERP
to language models.

How sales, service, and ERP data is connected to language models through the Model Context Protocol: GDPR-compliant and with an evidence trail. Drawn from the experience of the SIMOSphere AI Lab.

Connected systems

The Lab has connected seven systems. Each has its own MCP service that translates reads and writes into the protocol and logs every call.

Connected systems with category, authentication, and MCP service
CategoryAuthenticationMCP serviceScope
SalesforceCRMOAuth 2.0mcp-salesforceSales Cloud, Service Cloud, Marketing Cloud
HubSpotCRMOAuth 2.0mcp-hubspotContacts, deals, tickets, Marketing Hub
Microsoft Dynamics 365CRM and ERPOAuth 2.0 via Microsoft Entra ID (formerly Azure AD)mcp-dynamics365Sales, Customer Service, Finance and Operations
SAP S/4HANAERPOAuth 2.0 or SAMLmcp-sapOData v2 and v4, BAPI bridge, cloud and self-hosted alike
OdooERPAPI keymcp-odooSales, inventory, accounting
Microsoft 365CollaborationOAuth 2.0 via Microsoft Entra ID (formerly Azure AD)mcp-ms365SharePoint, OneDrive, Outlook, Teams
NextcloudDocumentsApp passwordmcp-nextcloudFiles, Talk, Deck, hosted entirely in the EU

Four layers, one architecture

No point-to-point tangle. Every layer has exactly one job, and every layer can be replaced on its own.

  1. Data sources

    CRM, ERP, Microsoft 365, SharePoint and Nextcloud stay where they are. Nothing is copied, nothing migrates into a second store.

  2. MCP services

    One service per system. It translates reads and writes into the standardized protocol and writes a log entry for every call.

  3. Orchestration

    It routes a request to the right model and to the right connector. The choice is Apertus, Mistral, or a provider using your own key. Permissions and data classification are enforced here.

  4. Workspace and roles

    Business users work in their familiar interface or with AI roles that have a clear scope. Nobody who would rather not know about MCP has to.

Time to value

Figures from completed client projects between the fourth quarter of 2025 and the second quarter of 2026. Elaborate custom fields or bridges to legacy systems take longer.

Measured setup effort per system
DurationWhat decides it
HubSpot and Odoo2 to 4 hoursAuthentication and a sandbox are ready within hours
SalesforceHalf a day to a dayProvided profiles and permissions are kept clean
Dynamics 365Half a day to a dayMicrosoft Entra ID and the tenant setup
SAP S/4HANA2 to 5 daysThe OData service and the permission model
Microsoft 365 and SharePointHalf a dayThe Graph API and the sites scope
Nextcloud1 to 2 hoursAn app password is enough

Consulting by SIMO GmbH

From the Lab to consulting.

SIMO does not judge options on vendor claims alone. Assumptions that can be tested technically are examined in the SIMOSphere AI Lab. SIMO GmbH works with your executive team to assess whether an approach that proves viable in the Lab makes sense for your company, and supports the implementation: independent of vendors, with a fixed scope and a decision after every phase.

Decision Readiness Check: 12 questions, about 3 minutes, instant result. Initial call: 45 minutes, free of charge, by video. Engagements begin with the BISA baseline assessment or the BEIA enterprise assessment. All on simo-online.com.