Data Security and Privacy Policies for Whatfix AI features

Whatfix uses a combination of Generative AI models from Microsoft Azure, OpenAI, and Google Vertex AI to support all Whatfix AI features. This ensures versatility and robustness in our AI offerings.

Whatfix uses several interlinked AI models that work together to provide seamless and integrated functionalities.

The Whatfix Generative AI model is built using Python and Java programming languages, chosen for their performance, scalability, and flexibility in handling complex AI tasks.

No, Whatfix does not use your content or data to train the AI model. The model operates independently of your data.

Whatfix does not require your PII data to use AI features. However, if PII is included in the input data, the AI-generated response might also contain PII. Whatfix follows strict guidelines to ensure that the PII data is handled according to agreements. Whatfix scrubs PII information from input data wherever possible.

The Whatfix AI model’s accuracy is ensured through continuous monitoring and evaluation. Feature-specific improvements are implemented to address any inaccuracies, ensuring reliable and accurate outputs.

No, Whatfix does not use AI to generate biased or influential content. Our focus is on delivering neutral, data-driven insights that align with your needs and business requirements.

Whatfix hosts its AI models using private or dedicated instances of Azure OpenAI, deployed within Whatfix's cloud infrastructure on Microsoft Azure. These instances are deployed in the data center or region corresponding to your location, ensuring data residency compliance and minimizing latency.

For example, if the enterprise location is within the EU, all queries go to private instances of these models deployed in the EU region.

Data Retention: Whatfix follows a standard data retention policy, retaining data for up to two years after the termination of the customer relationship.

Data Deletion: You can request data deletion, and Whatfix commits to processing such requests within 30 days as per the data retention and deletion policies.

Whatfix uses robust security protocols to ensure that the AI system is resistant to attacks. Model inputs are processed within a secure cloud environment, and we maintain certifications against globally recognized compliance standards, such as ISO 27001 and SOC 2, to guarantee the integrity of our AI models and infrastructure.

Model Training: The Whatfix AI system is largely autonomous, performing actions such as query processing and answer generation with minimal human intervention.

Human Role: Human interaction is essential for quality control, monitoring, and improving AI outputs to maintain high standards and meet enterprise needs. Whatfix ensures that the AI remains in line with industry best practices.

Whatfix’s AI features are designed to assist you, not make decisions on your behalf. The AI provides results, suggestions, and answers based on your queries, but it always requires human supervision to validate and act upon the outputs. This ensures that human judgement remains central to all decision-making processes.

Yes, Whatfix stores queries and responses generated by the AI features to help you understand the queries asked by users and improve the features.