What is the difference between reverse ETL and ETL?
+
ETL moves data from business apps into the warehouse, for reporting and analysis. Reverse ETL moves the results back out, from the warehouse into the apps, so people act on them. ETL writes to tables the data team owns. Reverse ETL writes to records other teams use every day, so field ownership and never overwriting a person matter more.
Which systems does Tray.ai connect for reverse ETL?
+
Warehouses (Snowflake, Google BigQuery, Databricks, Amazon Redshift), CRM (Salesforce, HubSpot, Microsoft Dynamics 365), marketing and ads (Marketo, Braze, Iterable, Segment, Google Ads), sales engagement and customer success (Outreach, SalesLoft, Gainsight), support (Zendesk, Intercom) and Slack for alerts. Anything without a connector connects through its API.
Can one platform run both ETL and reverse ETL?
+
Yes. On Tray.ai the workflow that loads Salesforce into Snowflake and the workflow that writes scores back run in one place, with the same credentials, logging and error handling. You do not need a separate tool for each direction.
Should reverse ETL run in real time or on a schedule?
+
Most syncs run on a schedule, every few minutes to every hour, and write only the rows that changed. Signals a person must act on now, like a risk tier moving to high, can trigger as soon as the model finishes.
Where do AI agents fit in reverse ETL?
+
Agents need the same account context people do. The warehouse results reverse ETL writes to the CRM also give an agent a current view of each account, and the same field ownership rules decide what an agent may change.
How is the data kept secure?
+
Tray.ai is SOC 1 and SOC 2 Type 2 audited across the platform. Credentials are held by the platform rather than in scripts, access is set by role, and every run is logged, so a value on a CRM record can be traced to the run that wrote it.