Arlie uses generative AI and interfaces with third-party Large Language Models (LLM). The default LLMs are open AI models hosted in Microsoft Azure, ensuring that your data remains private and is not used to train base models. To generate responses, Arlie uses your questions in the Chatbot, Commvault product documentation, knowledge base articles, tickets submitted by other customers, and the status and error description of selected jobs.
Arlie Architecture
The Arlie architecture has a control flow where you control the data Arlie responds to with a question or selected job. In the data pipeline, the ingested data is reviewed and processed before being used as context for the generative AI model. This process ensures accuracy, removes Personally Identifiable Information (PII), and extracts only the relevant data and metadata, providing a high level of data privacy and security.
Data Usage
Arlie uses the following data for generating the responses:
-
Active Insights: Customer tickets, knowledge base articles, selected job status, and error description
-
API Code Assist: User question or query, Commvault API documentation in api.commvault.com
-
Arlie Chatbot: User question or query, Commvault product documentation, Commvault store description text
-
Custom Walk-Through: User question or query, Commvault product documentation
Note
Arlie does not use your data to train the generative AI models. Furthermore, Arlie does not utilize your backed-up data for any purpose.