Use least-privilege access to connected services
Provider privacy terms and processing locations vary. Enterprise procurement should examine the data-flow diagram, subprocessor list, retention schedule and application of contractual safeguards.
ELP GPT / SAFETY
Security and privacy architecture
Security is built around identity, encrypted transport, data separation and least-privilege actions, including when ELP GPT invokes a third-party model.
OUR APPROACH
Security & privacy: principles in practice
Separate tenant data and restrict privileged operations
ELP GPT connects proprietary engineering to responsible, task-specific customer outcomes.
Provide clear privacy information and user controls
ELP GPT connects proprietary engineering to responsible, task-specific customer outcomes.
THE ELP GPT PERSPECTIVE
Security & privacy — engineering and execution
This page describes designed controls, not a claim of SOC 2, ISO 27001 or other certification.
ELP GPT's application uses server-side credentials, scoped connected-app permissions and private customer data stores. Requests are evaluated against access policies before tools can perform sensitive operations.
Technology, ownership & transparency
ELP GPT is developed by ELP Ventures. ELP-owned software and model research are distinct from independently owned inference services, third-party libraries and intellectual property. Specifications, availability and comparative performance depend on the deployed configuration.
Research and safety reference frameworks
Research and governance methods can be assessed against public resources such as the NIST AI Risk Management Framework and the European Commission guidance on general-purpose AI. These references inform evaluation and do not mean ELP GPT is certified or endorsed.