Our agent infrastructure operates with a "human-in-the-loop" design framework. Agents gather data, cross reference, and draft, but they don’t execute final decisions independently. You review, adjust, and approve every RFI draft, submittal review, and daily log to ensure human accountability and control.

No. Your project documentation and financials are walled off securely within your isolated tenant and are never pooled with other customers. Procore maintains strict enterprise agreements with model providers, ensuring that your proprietary project data is never used to train third-party foundational models.

Yes. Procore AI is permission-aware by design, meaning it dynamically mirrors your exact, role-based platform permission structures. A user can only query, analyze, or automate workflows using data they are already authorized to view in Procore.

Out-of-the-Box agents are pre-built, specialized tools designed to run common construction tasks with little or no setup. Agent Studio (available in the Enterprise tier) allows your team to use natural language instructions to configure custom digital coworkers, defining skills, custom triggers, and automated schedules.