Privacy Policy

At Nimblersoft, we take your privacy and data security seriously. This Privacy Policy details our approach to handling data, specifically outlining how we safeguard the information collected through our generative AI interfaces.

Agentic RAG Data Safety

Our application utilizes Agentic Retrieval-Augmented Generation (RAG) to provide a tailored, interactive experience. To protect confidentiality, we employ rigorous control over data flow when communicating with Large Language Models (LLMs).

Local PII Masking

Before any data is transmitted to an LLM, we ensure that all Personally Identifiable Information (PII) is masked locally.

Our system uses Microsoft Presidio directly on our secured host infrastructure. This ensures that names, email addresses, phone numbers, organizational references, and other sensitive details are replaced with sanitized markers (e.g., <PERSON>, <EMAIL>). As a result, the external LLM never processes or retains any associated personal information.

Local Observability and Security

For observability, interaction improvement, and audit purposes, conversation histories and query telemetry are kept locally.

• Local SQLite Logging: We securely store our logs (such as query_logs) exclusively as local SQLite databases on our servers. This design isolates analytical data and ensures that interaction traces remain strictly under our control.
• Controlled Retention: Query logs and captured service context details are structured for operational observability and isolated explicitly from third-party distribution.

Purpose and Use of Data

We affirm our commitment to respectful data practices:

• Information captured (such as expressed needs and service interests) is strictly used for the purpose of direct contact.
• You will only receive communication pertinent to your inquiries and sessions.
• No Spam: We do not sell your data, use the data for secondary marketing efforts without consent, or participate in spam operations.

Your interactions with Nimblersoft AI reflect a shared trust, and our privacy architecture is crafted to honor it comprehensively.