1. Our Guiding Principle: Your Data is Your Data
At Cust, we leverage advanced Large Language Models (LLMs) to power our AI features. We treat data privacy and security as a core component of our architecture, not an afterthought. Our platform is designed with a fundamental principle: your data is used exclusively to service your requests and is never sold, never used to train foundation models, and never co-mingled with data from other customers.
This document outlines the technical and contractual safeguards we have in place to guarantee strict data isolation and privacy.
2. Single-Gateway LLM Architecture
All LLM processing on the Cust platform (model inference and embeddings) runs through a single gateway: the Google Gemini Enterprise Agent Platform, operated under an enterprise agreement on Google Cloud Platform. Where third-party models are used, they are consumed as managed models on the platform, under the same contractual and technical controls. No customer data is ever sent to a consumer-grade AI tool or to any model provider outside this gateway.
A single gateway means a single set of data-protection guarantees to audit: one provider contract, one retention configuration, one access-control regime.
3. Zero Data Retention & No-Training Policy
This is our most critical commitment regarding your data.
- No training, contractually enforced: Google's Service Specific Terms contain an express Training Restriction: customer data is not used to train or fine-tune any AI/ML models without prior permission or instruction. This applies to all managed models on the platform, including GA and pre-GA models. Processing is further governed by Google's Cloud Data Processing Addendum (CDPA).
- Zero data retention: The platform is operated under a zero-data-retention configuration: prompts and outputs are processed ephemerally to serve the request and are not persisted by the model provider.
- All data processing is transactional: each request is a single, stateless exchange.
4. Our Secure Multi-Tenant Architecture
The Cust platform is built on a secure, multi-tenant architecture that ensures your data is isolated at every stage.
- Step 1: Strict data segregation in our database. All customer data is stored in our production database and logically segregated using a unique tenant ID. Every table and data record is tagged with this ID, ensuring a foundational layer of separation before any data is used in an AI-powered workflow.
- Step 2: In-memory prompt construction. When an authenticated user makes a request, our application retrieves only the necessary, tenant-specific data from our secure database and constructs the model prompt in-memory, on the fly.
- Step 3: Stateless gateway transaction. The in-memory prompt is sent to the Gemini Enterprise Agent Platform over an encrypted TLS 1.2+ connection for a single, stateless transaction. The response is returned immediately to our application.
This closed-loop process ensures each customer's data lives within its own secure, isolated context, used only for the immediate task at hand.
5. Platform Security Controls
We build on the enterprise-grade controls of Google Cloud Platform:
- IAM integration: All access to the Gemini Enterprise Agent Platform is governed by Google Cloud IAM with least-privilege service accounts; no long-lived personal credentials touch the inference path.
- Data residency: Google Cloud allows data residency to be enforced, ensuring data processing occurs within a specified geographic region. Regional (EU) inference is available as a per-tenant configuration.
- Compliance: Google Cloud Platform maintains a wide range of certifications, including SOC 2/3, ISO 27001, and HIPAA. Cust itself holds a SOC 2 Type II attestation covering Security, Confidentiality, and Availability.
6. Cust's SOC 2 Type 2 Compliance
Cust is SOC 2 Type 2 certified. This certification - independently verified by a third-party auditor - covers our security, availability, and confidentiality controls, ensuring we meet rigorous standards for protecting your data. Our SOC 2 Type 2 report is available to customers and prospects upon request.