MAG OptiAI Policies
Data Processing & Retention
This policy explains how MAG OptiAI processes and retains product data in the Pro Console.
It describes the main categories of product data, service providers, AI processing, retention windows, deletion behavior, and customer request paths.
Product data we process
Product data may include uploaded files, CSV rows, PDFs, text documents, operational datasets, scenario settings, constraints, resources, schedules, routes, extraction targets, model configuration, prompts, conversation turns, model outputs, AI explanations, and generated artifacts.
MAG OptiAI processes this data to provide the requested workspace functionality, preserve saved work, generate results, support product limits, provide billing/usage records, and troubleshoot customer requests.
Storage and retrieval systems
The backend stores account, project, scenario, model, document, conversation, billing, usage, and retention metadata in MAG OptiAI-controlled backend systems.
Compute artifacts and model/document result artifacts may be stored in Google Cloud Storage style artifact locations when configured. DocAI document retrieval uses Qdrant-backed vector storage for indexed document chunks and retrieval metadata.
AI and model processing
Supported AI features may send prompts, document text, retrieved context, user questions, result summaries, or structured inputs to AI providers such as OpenAI to generate answers, embeddings, extraction support, explanations, or comparisons.
Forecasting and fraud/risk workflows may create saved model versions and associated artifacts. Anomaly, extraction, DocAI, scheduling, routing, and other workflows may create saved result records, conversation turns, vector indexes, or result artifacts.
Current retention windows
For free-tier accounts, active unused saved work may be archived after 30 days. Archived saved work may be deleted after 7 days.
For paid or test accounts, active unused saved work may be archived after 90 days. Archived saved work may be deleted after 30 days.
Current projects, active models, protected project state, and records needed for billing, security, support, audit, legal, or service-integrity purposes may be excluded from automated cleanup while they remain needed.
Deletion behavior
Some product deletion paths mark records as deleted, archive records, or make records unavailable from active product views. Depending on the product workflow, related files, vectors, conversation data, model artifacts, document indexes, or artifact references may be removed, marked for deletion, or queued for cleanup by product or maintenance workflows.
Billing records, usage ledgers, security logs, support records, audit records, and records needed for legal, tax, dispute, abuse-prevention, or service-integrity purposes may be retained longer where required or appropriate.
Customer requests
Customers can contact MAG OptiAI for privacy, export, deletion, billing, or support questions. Requests are reviewed against account ownership, technical feasibility, product state, legal obligations, and current commercial arrangements.