Our Model Integration Policy
A common question is about security when using different Artificial Intelligence models available in Tess AI.
Our policy is clear: privacy is non-negotiable. The full security and data privacy that Tess offers is one of the main factors considered by our customers, especially companies.
Certification and Security: We only integrate into our platform AI models that have recognized security certifications, such as SOC 2, and that contractually guarantee data privacy. Models that do not meet these strict criteria are not included in Tess AI.
Additional Security: Using an AI model, like DeepSeek, through Tess AI is safer than using it directly (in its free version). We make sure the models run in our own infrastructure in a secure way, without sharing your data with the model providers.
Enterprise-Level Contracts
For models from large providers, like OpenAI’s ChatGPT, we operate with enterprise-level contracts (Enterprise). This means your data is protected by the same security and confidentiality agreements that protect large global corporations. Any interaction via API is governed by these terms, ensuring that your information is not used for training the models.
Secure Data Analysis with Deep Analysis
The Deep Analysis tool available in the chat and in agent creation was developed to enable analysis of large volumes of data without compromising security or accuracy.
In the chat:
In the Agent
How does it work?
When you use Deep Analysis, the AI is connected to a secure virtual machine (Tess Computer) that uses Python libraries to perform complex calculations and data processing.
The AI acts as an interpreter after this processing, communicating accurate results back to you without “hallucinations,” since the calculations are done by a deterministic computational environment.
Privacy Guaranteed: You can drag spreadsheets and documents with sensitive data into the tool, confident that the information will remain safe and private, being used only for the requested analysis.
