Harvey AI

Description Features Permissions Security & Compliance

Harvey for Slack brings Harvey, the AI assistant for legal professionals, into the conversations where in-house legal teams already work, so you can answer a question, add a key document to a vault, or review a document all without leaving the Slack thread.With Harvey in your Slack workspace, teams can @mention Harvey to ask legal questions, reason over documents. Harvey will quickly reply in-thread, with a link to continue longer work in Harvey.Harvey uses AI to generate responses, which may be inaccurate or incomplete. Review all output before relying on it, especially for legal work. Harvey for Slack requires an active Harvey subscription. A Harvey admin enables the integration and installs the app; each user links their Slack account to a Harvey account with a matching email

Harvey AI will be able to view:

Harvey AI will be able to do:

Review the details to better understand this app’s security practices. To learn more about assessing apps for your workspace visit our Help Center.

General

Privacy & data governance

Data retention policy

Customers can set retention policies and delete data at any time. Within 30 days after termination, Harvey securely deletes remaining Customer Data or Content unless the customer instructs otherwise. Customer Personal Data governed by Harvey’s Data Processing Addendum is returned or deleted within 30 days of service termination or a reasonable request, unless applicable law requires retention. Audit logs are retained for at least one year and up to ten years.

Data archiving and removal policy

Customers can set retention policies and delete data at any time. Within 30 days after termination, Harvey securely deletes remaining Customer Data or Content unless the customer instructs otherwise. Customer Personal Data governed by Harvey’s Data Processing Addendum is returned or deleted within 30 days of termination or a reasonable request, unless applicable law requires retention.

Data storage policy

Harvey stores and processes Customer Data and Content in the geographic region specified in the customer’s order form or otherwise agreed in writing. Data is logically separated by customer and encrypted using AES-256 or better at rest and TLS 1.2 or better in transit.

Data hosting company

Microsoft Azure

App/service has sub-processors

yes

Guidelines for sub-processors

https://www.harvey.ai/legal/subprocessors

App/service uses large language models (LLM)

yes

LLM model(s) used

OpenAI GPT-5, o3, GPT-4.1 and GPT-4o; Anthropic Claude Sonnet and Opus 4; and Google Gemini 2.5 Pro. Harvey selects models based on each request.

LLM retention settings

For Harvey’s core service, model providers use Zero Data Retention and may not retain data for human review. Optional admin-enabled Extended AI Features may have different retention terms disclosed in Harvey’s Service Terms.

LLM data tenancy policy

Customer workspaces are logically separated. Harvey and its model providers do not use Customer Data or Content to train shared models, and providers are contractually barred from using it to train or improve their services.

LLM data residency policy

Customer Data and Content are processed in the customer’s agreed region. Supported regions include the US, EU/Switzerland, and Australia. Model-provider locations are listed at harvey.ai/legal/subprocessors.

Certifications & compliance

Security