At Knak, we are committed to harnessing the power of AI to enhance our email and landing page creation platform, enabling our users to efficiently design and execute marketing campaigns. Our AI features are designed to seamlessly integrate with Knak, supporting our users directly within our platform. As we pursue our AI roadmap, we prioritize maintaining the highest standards of user privacy and data security, and we are committed to being transparent with our customers about the AI technologies we deploy, and provide the appropriate controls to tailor the types of AI used in their Knak instance.

We have chosen Amazon Web Services (AWS), Google Cloud Platform (GCP), and Open AI as our trusted partners for hosting our AI models. Our AI features operate on our private model instances hosted on these platforms. Importantly, our hosting partners do not use any of our users’ data to train AI models. We will provide notice to our customers who utilize the AI features in advance if we change our AI service providers.

While Knak may store certain data temporarily to enhance app performance or for identifying potential defects, this information is always encrypted and managed according to our stringent security protocols. Our commitment to data security ensures that user information is protected with the highest level of confidentiality and integrity.

Administrators have precise control over which AI features are activated via the settings page of Knak. They have the autonomy to opt out of any AI functionality at any time.

Additionally, to ensure transparency, we commit to notifying our clients whenever a new AI feature is introduced. This allows them to stay informed and make decisions that best suit their needs and preferences.

Our AI features undergo regular quality assurance (QA) for their performance, safety and bias. The QA evaluations follow our strict security protocols to maintain our users' data security and privacy. Should an issue be detected and can be resolved, improvements are applied and the model instance is updated accordingly. Otherwise, the AI feature may be deactivated while the issue is still under investigation.

We currently do not train or fine-tune our AI models using user data. Looking ahead, we may provide users the option to leverage their own assets to fine-tune models, but consistent with our commitment to user autonomy, participation in this feature will be entirely optional. Users will have the flexibility to opt in or out at any time, ensuring they maintain control over how their data is used.

We uphold the highest security controls and practices throughout the development and enhancement of our AI features. As we continue to innovate, our commitment to security remains steadfast, ensuring that both existing and new AI functionalities adhere to the rigorous standards our clients have come to expect from us.

In alignment with the broader Knak platform, our AI features are:

Knak is committed to maintaining the confidentiality of personally identifiable information (PII). We ensure that PII data is kept separate from other application data. We offer tools to help our users design marketing campaigns without using PII, including dynamic content allowing for personalization to be incorporated just-in-time during email dispatch. We extend this rigorous approach to our AI functionalities, which are specifically engineered to process anonymized content and do not require access to any PII data. We rigorously test and regularly update our AI functionality to ensure the highest data protection standards are met and in order to prevent any unauthorized access to personally identifiable information.

For conversational features, Knak temporarily stores individual conversations between the user and the AI to enable a continuous interaction. These conversations are handled with the same stringent data processing and storage protocols applied throughout our platform. Knak is SOC 2 Type II compliant and undergoes regular audits and pentests to ensure continued adherence to the highest security standards.

No. Customer data remains strictly private and is never shared with anyone. Knak does not train AI models. Our AI features are powered by our hosting partners (Amazon Web Services, Google Cloud, and Open AI), who do not use our data or our customers’ data for training or fine-tuning AI models. Knak has explicitly opted out of pre-release programs to ensure that neither our data nor our users' data is accessed for product improvements.

Knak encrypts all communications between its services including communications between our applications and users’ browsers (HTTPS). All data at rest is encrypted using AES 256 encryption and in transit via SSL/TLS 1.2 or higher.

Knak features comprehensive logging capabilities designed to detect errors and suspicious activities effectively. We take a proactive approach to logging, with numerous alerts configured for early detection of potential issues. All logging data is encrypted both in transit and at rest. Access to logging data is controlled and restricted to authorized employees only.

When a new AI model version is released by one of our hosting partners, we rigorously evaluate whether to proceed with an upgrade. This evaluation, conducted for each AI feature, includes both automated and manual testing to ensure continued functionality, security, and unbiased/quality results. We also re-evaluate the effectiveness of the automated safety and