What is Slite AI?
Slite (AI) Agent is a suite of AI capabilities built into your workspace. The Slite Agent lets you:
- Ask questions and find answers grounded in your own content, with source citations
- Connect data sources so the Agent can search across Slite, Slack, Google Drive, Notion, Confluence, Jira, Linear, HubSpot, Salesforce, GitHub, and more
- Build reusable workflows with Assistants tailored to specific tasks and sources
- Automate recurring updates with Digests sent to your team on a schedule
- Trigger actions with Contextual Buttons in the tools you already use
- Use the Agent in Slack and your browser via the Chrome extension
It's powered by trusted third-party model providers (Anthropic Claude and Google Gemini), all reviewed by our security team before use.
How does Slite AI work?
Which model providers does Slite use?
Slite AI runs on large language models from Anthropic (Claude) and Google (Gemini). Both are EU-hosted. We continuously evaluate providers to give you the best experience, and any third party that processes customer data is listed on List of Sub-Processors .
What gets sent to the model?
Only the snippets of content needed to answer your specific request are sent to the model, over encrypted connections. We use Retrieval-Augmented Generation (RAG): the AI searches your own indexed content, retrieves the relevant passages, and generates an answer based only on what it finds. It does not memorize or learn from your documents.
How is my data protected?
Will my data be used to train any models?
No. Slite does not use your content to train or fine-tune AI models — neither our own nor those of our providers. This applies to all customer data, including documents, queries, and generated answers, in any form. We have contractual agreements with our providers that prohibit them from training on your data.
Do your providers retain my data?
No. Our AI providers operate under a zero-retention policy: content is processed to answer your request and then immediately discarded. Nothing is stored, retained, or learned after a request is complete.
Does Slite AI respect existing permissions?
How permissions work depends on each source's access model:
- Mirrored sources match each user to their account in the connected tool by email, so they can only query content they already have access to there. Most sources work this way — including Slite, Slack, Google Drive, Confluence, Jira, Linear, and Salesforce.
- Shared sources are queryable by the workspace members an admin grants access to, regardless of individual permissions in the original tool. This applies to sources where per-item permissions aren't available, such as GitHub, HubSpot, uploaded files, and websites. By default a new shared source is restricted to admins, and access is configured per connection.
Because of this, admins should be mindful of which sources they connect and who they grant access to. See Agent Data Sources for the access model of each source and how to control it.
Is my data kept separate from other customers'?
Yes. Customer data is segregated, and we never mix or process data from different customers together during AI processing.
Where is my data hosted?
AI processing runs on EU-hosted models, so your data never leaves the EU. For full details on data hosting, residency, and encryption, see Security at Slite .
What compliance standards does Slite AI meet?
Slite is SOC 2 Type II certified and GDPR compliant. For more detail, see our Security at Slite , Privacy at Slite and List of Sub-Processors .
