
AI
OpenAI
AI workflows with source context and human review
OpenAI provides models used by Tali workflows, while Gratona governs the data and review path around them. Only the permitted context needed for the task is sent rather than an unrestricted record set. Tali-prepared work keeps source context attached so a reviewer can verify the facts, edit the draft, and decide what should happen next. Donor-facing or external action waits for human approval. OpenAI states that API inputs and outputs are not used to train its models by default unless the customer explicitly opts in; applicable retention still depends on the API endpoint and the organization's data-control settings.
Getting started
- 1Review the data boundaries and what is sent to OpenAI
- 2Set which records and fields are in scope
- 3Require human review before any output is sent
- 4Confirm audit logging and run a test draft
Key features
Defined data boundaries on what is sent
Source context on Tali-prepared work
Human review before donor-facing or external action
Review and audit context for Tali workflows
Frequently asked questions
Ready to connect OpenAI?
Set up OpenAI in minutes and start automating your workflow.