Does Ovida learn from customer conversations?

No. Ovida's AI does not learn from your conversations.

With Ovida, your conversations are yours. They generate insights for you, not training data for us. Our AI is pre-trained and never updated using your data.

This is a fundamental architectural decision and one of the things that makes Ovida safer than public AI tools for sensitive conversations.

Why this matters

Many public large language models (LLMs) continuously learn from user inputs. While this can make them more capable over time, it also creates real risks for organisations:

• Data leakage. Sensitive information entered into a public AI may end up in future training data, potentially exposing it to others.

• Loss of control. Once data is used for training, there's no way to remove it.

• Compliance concerns. Using customer conversations in model training can violate privacy laws like GDPR and HIPAA.

Ovida's approach

Our architecture is built differently:

• Models are pre-trained. They're built in advance and not updated with your data.

• Conversations stay private. Meeting recordings and transcripts are only used to generate feedback for you, never to train or expand the AI.

• Risk is eliminated. Because no customer data is used for training, the primary risks of public LLMs don't apply.

For more on how Ovida compares to public AI tools generally, see "Why is Ovida safer than using public AI tools?"