AskTable Data Privacy

Data types

In AskTable, the main data we process and store includes:

• •Chat history: conversations between users and the AI, including questions, answers, and interaction details.
• •Metadata (table and column names): structural information that helps the model understand how data is organized.
• •Chain-of-thought logs: reasoning steps used to improve analysis quality and answer accuracy.
• •Enterprise data: private business data that remains in your own databases.

Where data lives

AskTable offers two storage locations:

• •Cloud: hosted for accessibility, management, backup, and elastic compute. We use leading cloud providers to ensure security and reliability.
• •Local: data stays on your servers or private cloud, fully under your control for maximum privacy.

Privacy safeguards by deployment mode

1. SaaS

Your private business data stays local and is not uploaded to the cloud; chat history, metadata, and agent chain-of-thought logs are stored in the cloud. The LLM performs inference only and does not retain your data.

• •Encryption in transit and at rest.
• •Strict access control through authentication and authorization.

2. Standalone free edition

Private data, chat history, and metadata are stored and processed locally; chain-of-thought logs may be stored in the cloud for product improvement, while cloud models are used for inference (without retaining your data).

3. Private deployment (enterprise) A: cloud models

All data (business data, chats, metadata, chain-of-thought logs) stays on premises. Only cloud inference is used (no data retention by the model). Optional masked or encrypted inference sends de-identified values (e.g. names, phone numbers, IDs) to the cloud for compute.

4. Private deployment (enterprise) B: local models

Maximum privacy: all data is stored, processed, and inferred locally—ideal for the strictest security requirements.

Summary

Across these modes, AskTable offers flexible choices so you can match privacy needs to your environment. In every mode, we apply strong security and privacy practices to protect enterprise data.