Describe the workflow

Write what the agent should do, in plain English. Include what it reads, what it produces, who reviews it, and what counts as a good outcome. ownEvo will generate a simulator, an eval set, and a success metric from your description — you'll review and adjust them on the next step.

Already have an agent running? Skip the description — connect your existing traces and we'll infer the workflow. Connect existing agent →

Add a data source Add a success metric Specify a reviewer Add an example outcome

Workflow description 412 / 4,000 chars

You can edit the workflow description at any time. Each edit is versioned and gate-able like a skill change.

Generate sim & eval set →

What ownEvo generates next: a simulator (synthetic SKU catalog, supplier behaviour, weather + sales generators), 12 representative eval cases derived from your past misses + edge cases, and a success metric definition (markdown-window precision/recall + discount-bp accuracy). You review and edit each before the baseline run.

Or start from a template

Supply chain

Labour shift scheduling

Optimize weekly shift assignments under union OT caps and forecasted demand.

Legal / Labour

Union contract review

Flag risk clauses, jurisdictional carve-outs, and grievance precedent in proposed contract changes.

Customer ops

Customer support triage

Triage incoming tickets, draft responses, escalate refunds above policy threshold.