Every labour decision has a window. Early in the week it's an edit on the rota. By Friday it's a phone call to someone who has already made weekend plans. Slate is built around the early part.
Ask a manager how long a what-if takes. What if we cut Tuesday lunch by one person? Usually a calculator and a guess. Slate is being built to answer it straight away, with expected covers next to the hours.
A shift swap that would push someone over their hours should say so before anyone approves it. Slate is built to show that on the swap itself. The manager still presses yes.
Month end. The spreadsheet, the group chat and the notepad all have a different number for wages.
August came to £19,718. That's 29.9% of £66,000 turnover.
Slate is pre-launch. Features shown are previews. Figures are illustrative.
Big groups can pay someone to read labour against sales. An independent has the owner at 11pm with a spreadsheet. Slate is being built to do that reading every day and flag the odd bits.
The labour number in most restaurants arrives as a verdict. Slate is being built to make it a forecast: where this week is heading against plan, while there's still time to steer.
Building a rota from a blank grid eats an evening. Reading a draft and moving the handful of shifts that look wrong takes far less. Slate is being built to hand the manager the draft.
Rent won't change on a Thursday. Neither will the price of butter. A shift can. That's why Slate watches labour against plan while the week is still open, because labour is the big cost you can still move.
Most restaurant software records what happened. Slate is built to tell you what's about to happen. How busy Saturday looks, whether this week's hours fit it, which shift breaks a rule. Then a person decides.
That is the difference I'm building towards: a labour OS with an AI layer, not another clock-in screen. It spots the next decision and leaves the judgement with the person running the floor.
A rota app can publish the week. The useful bit starts before that: what demand looks like, what hours you've booked, and where the plan is already drifting.
The layer keeps checking as the week changes: cover against demand, hours against plan, rules against the rota. A gap is useful when the week is still editable.