Blissbook uses generative AI in two ways: as product features available to customers, and as internal tooling to help run our business. This article covers both, along with the controls, commitments, and policies that govern them.

Blissbook's AI features help you draft, summarize, suggest improvements, answer questions, and speed up tasks within the platform. Here's what you need to know.

By default, AI requests are processed within Blissbook's existing IT infrastructure, within the same environment that hosts all of your data, under the same security and privacy controls. We don't send your data to a separate third-party AI provider unless you choose to.

If you'd prefer to use your own AI provider (and you're on our PRO plan), you can enable Bring Your Own AI (BYOAI) and supply your own API key in AI Settings. When BYOAI is enabled, Blissbook routes AI requests to the provider you select. You're responsible for reviewing that provider's terms, privacy practices, and any compliance obligations that come with it.

AI features are managed by Org Admins through AI Settings.

The first time an Org Admin accesses an AI feature, a notice pops up summarizing how AI works at Blissbook, including security, privacy, and available controls.

We don't train AI models on your data. Your content, prompts, and AI outputs are not used to train Blissbook's or any third party's AI models.

AI prompts and outputs may be stored alongside your content within Blissbook and logged for troubleshooting, abuse prevention, and security purposes.

Blissbook's AI features are not designed to process your employees' personal data. The platform is built to exclude employee PII from AI processing. If personal data does enter the system (for example, if someone types it into a chat prompt), it stays within our existing infrastructure and is subject to our standard data handling practices, including automatic deletion of chat content after 30 days.

AI output may be inaccurate. It's generated probabilistically and should be reviewed by a human before you rely on it, publish it, or communicate it to employees.

AI output is not legal advice. Only your own attorney can determine whether the information, and your interpretation of it, applies to your particular situation. Contact your legal counsel for advice on any specific legal matter.

You own your content and your AI outputs. Blissbook retains rights to the platform and AI features themselves.

Blissbook uses Anthropic's Claude AI tools (Chat, Code, and Cowork) for internal productivity across the company. Claude is connected to several of our internal business tools via integrations.

Through these integrations, Claude may process customer business information such as company names, limited email addresses and email content, and metadata about your Blissbook subscription. Claude does not have general access to our customer's employee data or PII. Incidental exposure to small amounts of employee data is possible (for example, if a customer emails us an employee list). We don't intentionally provide such data to Claude and regularly delete it from our systems.

Per Anthropic's commercial terms, inputs and outputs are not used to train Anthropic's models, and inputs and outputs are automatically deleted within 30 days of receipt or generation.

For full details, refer to the following:

Note: These documents may not be updated yet as of 2026-4-9. They will be updated soon.

Common questions from IT, security, and privacy teams reviewing Blissbook's AI features.

Not today. If you have a specific need, such as jurisdictional restrictions tied to local AI laws, contact us so we can understand the use case.

For the Audience Chatbot, prompts and outputs are stored encrypted in our database so audience members / employees can resume recent conversations. A limited group of engineers with production database access and the encryption key can access them.

For the Audience Chatbot, yes. Conversations are tied to the audience member so they can see and resume their own chats. Chats automatically delete 30 days after the last message, and the audience member can delete them sooner. Customer Admins cannot view their employees' prompts or outputs.

Not through Blissbook's own AI features. Blissbook doesn't make automated decisions about employees. If your team builds on top of Blissbook (for example, pulling handbook data into your own HR decision system via a future API or integration), that logic lives in your system, not ours.

If you have questions about how Blissbook uses AI or want more detail about our infrastructure, contact us at support@blissbook.com.