CASE STUDIES
Supplement an external model and limit confidential data
A technical sales department uses AIDA locally for confidential documents. For general text drafts, an external model from OpenAI, Anthropic, or another approved provider should be added.
View larger image ↗
ILLUSTRATIVE CASE STUDY · NOT A CUSTOMER REFERENCE
The starting point
The text should be written in an easily understandable way, but does not require all construction details. Before integration, it must be clear which content will be transferred and which provider terms and conditions apply to the access being used.
This is how the process is intended to work
-
Separate task and data
The team describes the desired text and identifies confidential content that is not needed for this step.
-
Check integration with AIDA
The security audit considers the intended access, data volume, and processing method. The provider's current storage and usage terms are also taken into account.
-
Set permissions and restrictions
Those responsible determine the scope of data that may be used for the task. Technical controls and responsibilities are agreed upon for the specific connection.
-
Review the result through sales
The draft text will be reviewed by experts. Technical specifications, prices, and communication with the client will be approved by the responsible personnel.
The prepared result
External modeling capability can be used for a specific, defined task. Confidential sources are not automatically released for further external processing as a result.
Data and responsibility
This test applies to the described use case. A different model, new tools, or additional data may require a retest.
View larger image ↗