Labor targets in plain words: "keep Midtown at 28%" with ChatGPT or Claude

Most owners have a labor number in their head. "Midtown should run under 28%." "Riverside can go to 30% while it's new." The trouble is that a number in your head doesn't check anything. It doesn't warn you when next week's schedule is heavy, and it doesn't tell a new manager what you expect.

Mays keeps a labor target per schedule week: the labor card on the schedule. With Mays connected to ChatGPT or Claude, you can set it the way you'd say it out loud.

This post is part of our series Run your restaurant with your AI.

The situation

You run Harbor Lane Café's two locations. Midtown is established and should stay at 28% labor. Riverside opened recently and you'll accept 30% for the next month.

Before: the targets live in a spreadsheet, or a sticky note, or nowhere. Every Monday you export sales and labor, work out the percent, compare it with the number in your head, and message the managers. Next week's schedule is never checked against it until the week is over.

Now:

Keep Midtown at 28% labor for the next 8 weeks, and set Riverside to 30% for the next 4.

What the AI does in Mays

It finds both locations in your account.
It asks Mays to set the targets (set_location_targets): a labor % target for the week you start from and up to 12 weeks ahead. Mays sends a preview first, week by week: the target before and after. Nothing changes yet.
Your yes saves them. These are the same weekly targets you see on the labor card on the Mays schedule, not a separate copy that lives in the chat.

You can also set a weekly labor budget in dollars ("Riverside's labor budget is $6,500 a week"), if you can see pay in Mays and your connection includes pay. The labor % target doesn't need pay access to set.

To check what's saved: "What's Midtown's labor target this week?" (get_location_targets).

What the target does afterwards

Once a week has a target, the answers that touch labor check against it:

The schedule. Ask "How does next week look at Midtown?" and, with pay access, the answer says when the week as scheduled would go over the target: the scheduled labor cost against the sales forecast (get_schedule).
A suggested week. "Suggest next week at Riverside" runs Mays's auto-scheduler and compares the projected labor % with the location's own target, unless you name a different one in the question. See drafting next week against your forecast.
Last week's results. Sales and labor answers say when a location went over its labor % target.

All of it is advice. Mays never blocks a shift because a week is over target. Sometimes a heavy week is the right call (a catering order, a new hire training), and that's your decision.

Pay access and what you'll see

Labor percent is labor cost divided by sales, so the comparison needs wages. If you ticked Include wages, labor cost and tips when you connected, and you can see pay in Mays, the answers show the labor cost and the labor % against the target. If not, the target is still saved and still shown, but the AI can't say whether a week is over it, and it won't estimate.

The forecast behind "next week" averages the same weekday over the last 8 weeks and skips days with no sales, so it runs a little high, and labor % against it reads a little low. A week projected just under target is about on target.

What Mays doesn't do here

It doesn't block anything. A target is a line to compare against, not a cap.
It doesn't set targets by itself. You say the number; the AI never invents one. If you ask "What should my target be?", the AI can show you your recent labor % from Mays, but the target is yours to choose.
It doesn't add locations together. Each location has its own target, week by week.

Before you start

Connect Mays to your AI app. mays.app/ai has the setup for ChatGPT and Claude; Connecting Mays to your AI app has the details.
Reading targets is a read. Setting them is a change: an admin turns on Allow AI apps to take actions under Settings → Connected AI apps → Organization, and you tick make changes when you connect. Every change shows a preview and waits for your yes.

Once Midtown has a target, a midweek overtime check is the easiest way to protect it.

Try Mays free for 30 days, with no credit card, and connect it to the AI you already use.