GreeneStep engineered AI Agents follows enterprise-grade security protocols, ensuring that all uploaded data (PDFs, CSVs, XMLs, etc.) is securely stored and accessed only by authorized users. It includes model key checks, encryption, and access controls to protect sensitive business information. 

When you use an OpenAI API key or any paid API keys for generating responses, your input data is sent to OpenAI’s servers for processing. OpenAI handles data based on its privacy policies, and depending on your API settings, data might be logged temporarily.

For Company’s private and internal data, Greenetep AI Agents are trained internally on GreeneStep Servers. We use Ollama (Llama and other open-source models) for processing your data entirely within Greenestep’s secured servers. This means your private data never leaves our infrastructure, ensuring full security and compliance for confidential use cases. For Reporter and Trainer Agent, we use this model.

GreeneStep AI Agents are designed with privacy-first architecture, ensuring compliance with general protection standards. User confidential and private data is not shared, sold, or used for training external models, keeping company-specific information confidential.

We support Bring Your Own Token (BYOT) for AI models like OpenAI, Llama, and other leading providers. This allows businesses to have full control over API usage, cost management, and security preferences. You can use the Token keys update option in the application to use your own tokens. 

All end-user interactions with the AI agent are securely logged in private and secured servers of GreeneStep that are accessible only to authorized personnel. These logs help businesses track engagement while maintaining strict access control. 

GreeneStep AI Agents access includes multi-layer authentication, role-based access control (RBAC), and secure API integration and validations to prevent unauthorized access.