How Deal Brain treats your data
- Read-first by default. Deal Brain reads from your systems to build context. The writes it makes back to your CRM follow the update mode set on each field — see Fields and CRM objects.
- Your data is yours. We don’t train foundation models on it.
- Scoped access. We ask for the minimum permissions each connection needs. Scopes are documented on an integration’s page where they have been confirmed.
- Human in the loop. Outbound drafts never send automatically — you review and send every one. CRM writes depend on how each field is configured, and not every field requires review.
Where your data flows
Deal Brain uses the data synced through its integrations — emails, calendar events, call notes, and CRM records — in prompts to large language models (LLMs). This is how Deal Brain forms associations between data points and suggests updates to be made in your CRM.Models and providers
Deal Brain uses large language models from multiple providers to process the data synced through your integrations. We reevaluate which models we use as capabilities and costs change, so the specific set shifts over time. If your security team needs the current model list or our subprocessor list, get in touch and we’ll send it to you.Data collected by Deal Brain is not used to train models.
What each integration accesses
Permissions are scoped per connection. Each integration page covers what Deal Brain accesses and what it does with that access, and lists the specific OAuth scopes where they have been confirmed. If a page you need doesn’t list them, get in touch and we’ll confirm them for you.Salesforce
OAuth scopes and CRM read/write behavior.
HubSpot
What syncs and what Deal Brain does with the access.
Google Workspace
Gmail and Calendar read scopes.
Granola
Meeting and transcript access.
Slack
Channel-scoped ingest, including brain-emoji-only mode.