Slack AI Agents — privacy & data policy
Provider: Salesforce. PromptPrivacy score: 71/100. Last verified November 01, 2026.
What happens to your data?
Customer Data such as messages, files, and interactions are processed on behalf of the customer who manages the workspace. Slack also collects usage metadata, log files, and device identifiers to secure, maintain, and improve the platform. Information may be shared with vetted subprocessors, corporate affiliates, or legal authorities when required.
Does it train on your data?
Not specified. The policy does not state whether user content or inputs are used to train AI models; it only notes that Other Information (such as metadata and usage patterns) may be used to develop predictive models and feature suggestions under legitimate interests.
Can you opt out?
Not specified. The policy does not address AI model training on customer content or specify any opt-out mechanisms for AI processing.
Data retention
Customer Data is retained according to the customer's instructions and custom settings configured at the workspace, channel, or file level. Metadata and operational information are kept for as long as necessary to provide services, pursue legitimate business interests, comply with legal obligations, and resolve disputes.
Encryption & security
Slack states that it has received internationally recognised security certifications and takes steps to protect data against unauthorized access, loss, or disclosure. However, technical encryption protocols like TLS or AES at rest are not detailed within this text.
Is it safe for work?
Yes for general workplace collaboration, as Slack provides enterprise-level controls, a Data Processing Addendum, and customer-managed retention. However, because this policy does not contain terms specific to Slack AI Agents or model training on customer inputs, organizations should verify supplementary Slack AI terms before deploying sensitive data to AI agent features.
Source policy: https://slack.com/trust/privacy/privacy-policy