Human-confirmed · Deterministic · Explainable

Describe shift constraints in plain language — get atomic, auditable conditions.

ShiftPilot interprets Thai, English, or mixed scheduling requests into atomic condition cards. Nothing becomes an active rule until a human confirms it.

The operating principle

The language model interprets. A human confirms. The checker verifies.

In this proof of concept, an AI proposes the roster and a deterministic in-browser checker re-verifies every hard rule, labelling any unmet soft preference.

How it works

Four steps from plain language to a roster the checker verifies.

Live interpretation demo

Describe a scheduling request. The interpreter returns atomic condition cards for you to confirm.

Try an example:

More synthetic examples →

Team shift schedule

Add each person's confirmed conditions, set the shifts, then let the AI propose a roster — checked again deterministically in your browser.

Team

No one added yet — interpret a request, confirm the cards, then add the person.

Shifts

Morning 08:00–16:00
Evening 16:00–00:00
Night 00:00–08:00
Starting Tue 1 September 2026 (Asia/Bangkok)

Built for real scheduling

Nine capabilities that keep scheduling fair, explainable, and auditable.

Any industry that runs on shifts

If your operation runs on shifts, ShiftPilot fits.

Four deployment postures

The same product, your control — from public cloud to fully air-gapped.

Ready to pilot ShiftPilot?

Run the demo above with your own requests — or talk to us about a pilot in your industry.

Try the demo

Proof of concept. No accounts. No data leaves your browser except the interpretation request and the schedule request.