From policy to approval route
Describe the policy in one sentence. Review the draft, approve it, and it is live. No ticket, no chasing.
You describe it. We draft it. You approve it.
“Leave over 5 days needs the manager.” That is the whole brief. Nothing to learn, no ticket to file, no diagram to draw yourself.
The route appears with every approver in place. Fix anything with a click, then publish. Nothing goes live without your sign-off.
Every request finds its approver on its own. You always see where things stand. Nothing stuck, nothing lost, nobody to nudge.
In build. Three places, and nowhere else.
HR policy already lives in documents and intranet pages. AI retrieves the rule that applies to the process the manager describes.
It returns a workflow draft: states, conditions, approval chain, and the edge cases the document mentions. The manager edits it on the canvas.
Nothing publishes itself. The manager reviews the draft, fixes it, and approves it before it becomes a live route.
Model choice. We build on Claude. It is the only model in our AI layer, and it is not a placeholder we plan to replace later.
The same engine reads every domain. Only the sentence changes.
Each module ships its own states, fields, and conditions.
Annual, medical, and parental requests with balance fields.
Claim review, budget thresholds, and finance queues.
Offer stages, approval gates, and rejection paths.
Task sequences across IT, payroll, and the hiring manager.
Review cycles with self, manager, and calibration steps.
We would rather you read this than find out on a call.
We run the product on a call, from an empty canvas to a published route. If it is not a fit for your HR team, you will know in ten minutes. Write to [email protected].