Setting Up System Services
Status: 03.09.2025
On the Configure Services page, you set up all necessary AI services required for using Smart Processing. These services form the foundation for AI-supported functions in Smart Processing. The service connections (Service Connections) are set up individually by the customer and can then be used throughout the Simova Core ecosystem.
Note
New services are not provided by Simova. The setup is done directly by the customer through the respective provider, such as Azure. Existing services can be maintained or new ones created via the Simova Apps | CORE - AI Service List page in Simova Core.
Available AI Services
The following services can be integrated into Smart Processing:
- Azure OpenAI Service
- Document Intelligence
- Content Understanding
Setting Up Services
The setup of the services is done directly through the Smart Process Setup.
Step-by-Step Guide
- Start in the Smart Process Setup area.
- Select the action Configure Services.
- Check if a service already exists. If not, click on Manage OpenAI Service to create a new service.
- Save your entries. The configured services will then be available system-wide in the Simova Core ecosystem.
Functions and Associated Services
The following table shows which service is required for which function in Smart Processing:
| Function | Required Service |
|---|---|
| Document Intelligence Classification Method | Document Intelligence |
| Content Understanding Classification Method | Content Understanding |
| Document Intelligence Capture Method | Document Intelligence |
| Content Understanding Capture Method | Content Understanding |
| Option "Smart Field Matching" in Process Template | Azure OpenAI |
| Processing of XML Documents | Azure OpenAI |
| Processing of Email Documents (MSG, EML) | Azure OpenAI |
| Option "Allow Smart Master Data Creation" in Template Field | Azure OpenAI |
Use of Codeunit 5673302 "SIM_DI Smart Line Type Match" |
Azure OpenAI |
Recommended Models for Azure OpenAI
The choice of the appropriate model depends on the use case. The following models are recommended for most scenarios:
- gpt-5.2
- gpt-5.2-mini
Best Practice
Always use the latest model version. New models are generally more powerful and cost-effective.
Error Handling
If a required service is missing or the connection is faulty, Smart Processing automatically displays an error message when executing the action. The message describes which service is unreachable and what steps are necessary for correction.
Microsoft Foundry Service
AI-based functions are subject to so-called quotas. Quotas determine how many AI requests and how much data processing are possible within a certain period. If the available quota is reached, AI functions may be temporarily restricted or delayed. Depending on the use case, an increase in quotas can be requested from Microsoft to enable higher or more stable usage. For more information and the possibility to request, see Microsoft Foundry Service: Request for Quota Increase