Start in plain language or with explicit controls
Agents can be drafted from a prompt or built from explicit triggers, filters, logic steps, tools, and risk categories.
Personal AI automations - Beta
Automation with a visible boundary.
- Each agent runs as you, with your current access.
- Sensitive actions ask for a higher confirmation before they enable.
- Every run records what completed, failed, or was denied.
Illustrative product view. Sample data is fictional and permission labels are educational.
How it works
Concrete product states, permissions, and handoffs
Each demonstration uses fictional sample data and keeps the source record, actor, permission boundary, and next step visible.
Triggers that match the work
The tool catalog is not a wish list: selected tools run as the owner and are rechecked against live access.
A small, inspectable logic language
Sensitive actions require a visibly higher level of confirmation before enablement.
Tools are permissions, not a wish list
Agents can be drafted from a prompt or built from explicit triggers, filters, logic steps, tools, and risk categories.
Risk is explicit
The tool catalog is not a wish list: selected tools run as the owner and are rechecked against live access.
Validate before enabling
Sensitive actions require a visibly higher level of confirmation before enablement.
Product map
Follow the next useful path
These links are curated for the page rather than generated at random.
Workflow automation - PreviewTurn a people process into a flow everyone can follow.Bring forms, tasks, timing, documents, reminders, and approvals into one visible lifecycle flow.AI assistant - AvailableAsk the workspace. Act without losing control.Ask questions, create work, and move people operations forward with an AI assistant that follows Perelan permissions.Security at PerelanSecurity is part of the action path, not a page added after it.Understand how Perelan approaches authentication, authorization, service identity, integration security, AI boundaries, auditability, files, tokens, and incident readiness.Status & approvals - AvailableSee what needs attention, and what changed after you acted.Bring pending decisions, relevant notifications, and configurable widgets into one focused attention layer.
Questions
What buyers usually ask
Answers are scoped to repository-verified behavior or intentionally point to current customer documentation when policy facts are not public.
No. Agents act as the owner and use the owner's current access on future runs.
The product model includes manual and scheduled/event-style triggers. Availability depends on the configured agent and owner permissions.
Run records expose completed, failed, and denied states so the owner can inspect what happened.
Agents are marked Beta in the product and on this site.
No. The marketing examples are compressed, but the product path still checks the acting user, role, group scope, and domain permission before data is shown or changed.
No. The product demonstrations use fictional names, roles, dates, and values so the workflow is clear without exposing real people or companies.
Next step
Build a people operation your team will actually use.
Start with the workflow you want to make calmer, then map the people, records, permissions, and handoffs that make it real.