Data & AI usage
This page describes how customer data interacts with the AI components of Pincera. It should be read alongside the Privacy Policy and Security page. Specific commitments for an individual customer are set out in that customer’s ordering documentation.
Our position
Enterprise customers expect clear answers about how their data is used in combination with AI. The commitments below apply by default. Where a customer has agreed to different terms, those terms control.
- We do not use customer data to train third-party generative AI models. Customer prompts, responses, and business data are not shared with model providers for training purposes.
- We do not use customer data to train Pincera to serve other customers. Models, prompts, and configuration are scoped to each customer’s deployment.
- Customer data is not sold or shared for advertising.
How AI is used inside Pincera
Pincera uses AI to interpret a user’s request in natural language, to assemble the right context for a task, and to help complete that task against the systems your business already runs. AI operates inside the limits set by your configuration — the apps connected, the rules defined, and the approvals required before action is taken.
Data sent to model providers
To interpret a request, Pincera may send limited, task-scoped input to an AI model. This is the data needed to perform the specific request — not a broader export of customer data.
- Inputs are sent under data-processing terms that prohibit the provider from using them to train or improve their models.
- Inputs are sent with a zero-retention commitment where the provider supports it. Where retention is unavoidable, we minimize it to what is required to operate the service.
- We maintain a current list of AI sub-processors. Customers and prospective customers under NDA can request it through their account contact.
Isolation between customers
Each customer’s deployment is isolated. Prompts, data, and application state are scoped to the customer. Pincera does not combine customer data across accounts to improve a shared model.
Data location
Pincera supports deployment regions that match common enterprise data residency requirements. Specific regions and residency guarantees are agreed with each customer in their ordering documentation.
Retention and deletion
Customer data is retained as agreed in the customer’s ordering documentation. On termination or on request, customer data is deleted or returned in accordance with that agreement. Inputs sent to model providers follow the retention terms described above.
Audit and oversight
Every Pincera action on a customer system is recorded in an audit trail the customer controls. Customers can review which requests were handled by the Service, who approved sensitive actions, and what data was touched.
Model choice and updates
The specific AI models used to serve a customer may change over time as we adopt improved versions or alternate providers. Where a customer has specified a required provider or model profile, that choice is honored under their ordering documentation.
Human review
Pincera personnel access customer data only when required to support or secure the Service, subject to approval and logging. Customer data is not reviewed by humans for the purpose of improving a model unless the customer has explicitly agreed to such a program.
Requests and questions
Customers can request details on data handling, sub-processors, and region-specific arrangements through their account contact. General inquiries can be sent to info@nacsoftech.com.