Construction scheduling software for general contractors running 3-6 active jobsites. AI scheduling for GCs and multi-jobsite scheduling that handles the way you actually run weather days, MEP rough-in dependencies, OSHA-cert role gating, multi-site foreman coordination, and the end-of-month OT push. $200,000 to $600,000 a year per GC of preventable bleed. Caught at save time, not at payroll close.

Walsh, your superintendent, used to find out about overtime when payroll closed the next Friday. By then the premium was already booked, the foremen were defensive, and the only conversation was "we'll watch it next week." On Midtown Tower, 11 staffers were already over 40 hours by Tuesday morning. At a framer base rate of $24-$28/hr and a journeyman electrician base of $34-$42/hr, the 1.5x overtime premium per staffer per OT hour runs $12-$21/hr. If Walsh does not intervene by Tuesday afternoon, the average over-40 staffer on Midtown lands 6-10 OT hours on the week. 11 staffers × 8 OT hours × $16 average premium = $1,408 of OT premium on a single jobsite for a single week. That is one week. One jobsite. Workforce Insights surfaces scheduled hours and OT trending per staffer, per role, per jobsite, week by week. Walsh now opens the dashboard every Tuesday morning and sees the trend before the bleed prints. He moves the framing crew on Midtown to Park East where the foreman is light, calls the crane operator's apprentice to cover the back half of Wednesday, and the week closes 7 OT hours instead of 88. Across 4 active jobsites a 3-6 jobsite GC saves the OT premium 30-50 times a year — not because the OT vanishes, but because Walsh catches it on Tuesday instead of next Friday.

Park, your project manager, sets a weekly labor budget per jobsite — Midtown Tower at $38,000, Park East at $26,500, Westside at $14,000, River Bend at $11,000. With manual scheduling and legacy construction workforce platforms, the budget is a number on a spreadsheet. Foremen schedule whatever crews they think they need. The real labor cost lands at payroll on Friday. By then the variance is whatever it is. On River Bend, the foreman scheduled two extra ironworkers Thursday for a steel lift that did not happen. $3,000 over on a single week. Across 50 weeks × 4 active jobsites × an average $400-$1,500 of preventable variance per jobsite per week = $80,000-$300,000 a year on a 4-jobsite GC. In XShift, on the Autopilot page you write a custom rule in plain English: "weekly labor cost over $11,000 at River Bend — block the shift." Add the same rule for the other 3 jobsites at their numbers. The rule fires when the shift is being assigned, not at payroll. A shift that would push the week over the ceiling either blocks at save, or shows a manager-override popup with a written-reason box. Same for the daily cap — "daily labor cost over $2,200 at Midtown on this day — block." The 5-rule cap forces Park to pick the budgets that matter most. But the labor budget stops being a Monday-morning surprise. It starts being a Friday ceiling the grid actually holds. Heads up: the labor-cost rules block or warn. They do not auto-pick a cheaper crew. That is the foreman's call.

XShift handles certs through roles. There is no cert-expiry date stored inside the scheduler. Instead, you add a role on each staffer's profile for every cert they hold. Roles like "signal person," "rigger," "long safety course done (super)," "short safety course done (carpenter)," "hot-work permit," "confined-space cert," "fall-protection trained," and "scaffold trained." Each cert is a role. Each staffer with the cert gets the role on their profile. When the cert renews or lapses in your HR system, you update the role. That is the enforcement gate. Then in the staffing rules, you set "Westside Office Fitout, Tuesday 7 AM-3 PM hot-work block: need 1 hot-work permit role on shift." The AI Copilot honors the staffing rule when it builds the week. Any save that would leave the hot-work block uncovered gets blocked at save time. This is not a compliance guarantee. XShift makes no compliance promises. What it does is hold the line on whoever you have set as eligible. The cert lapse itself still has to live in your HR system. The contract-clause exposure when a non-cert staffer ends up on a cert-required job is the reason the role-as-cert workaround exists — $10K-$50K per-incident penalty clauses are common in commercial GC contracts. The AI Copilot will not put a non-eligible staffer on the shift. The grid blocks it. Aldana, your foreman on Westside, sees the red tile, calls the journeyman who got his hot-work permit last month, swaps him in, the tile goes green.

Construction shift coverage is not a flat number across the day. The crane lift block on Tuesday morning at Midtown needs a crane operator plus 2 riggers plus a signal person — 4 roles, all needed, only for that 4-hour window. The Wednesday morning concrete pour at Park East needs a finisher plus 3 laborers plus a foreman, 5 AM to 9 AM. The Thursday MEP rough-in at Westside needs 2 electricians plus 2 plumbers plus an MEP tech, 7 AM to 3 PM. If any of these windows runs short, the critical path slips. Critical-path slip cost in commercial construction runs $5,000-$50,000 per day in carrying cost, equipment idle, and downstream-trade rebooking. Manual scheduling holds these minimums in Walsh's head. Conventional construction workforce platforms let you set a flat per-shift minimum but not a per-time-window minimum — and the older jobsite scheduling tools have no concept of trade-sequence coverage at all. In XShift, you set a per-window minimum: jobsite, role, day-of-week, time window, count. The rule fires when the shift's start and end match the rule's window exactly. The AI Copilot honors the rule when it builds the week. The grid blocks any save that would leave the window under the minimum. Heads up: the time match is exact. A 7:00 AM to 3:00 PM rule does not fire on a 7:00 AM to 3:30 PM shift. Set the rules to match the shift windows your crews actually run.

A GC running 3-6 active jobsites with separate spreadsheets is running 3-6 separate scheduling problems. They all overlap on the most expensive staffers — the foremen, the crane operators, the finish carpenters, the journeyman electricians, and the supers with the long safety course done. Foremen routinely get double-booked across two jobsites on the same Tuesday morning because nobody sees both sites at the same time until Friday. A double-booked foreman on a critical-path day costs $5,000-$50,000 in delay carrying. Across 8-15 double-bookings a year for a 4-jobsite GC: $40,000-$750,000 of avoidable critical-path slip. Multi-location is included in the XShift base tier — no paid add-on, no upgrade, no separate seat. One organization, every jobsite, one staffer roster, one schedule view. The AI Copilot sees every staffer's hours across every jobsite when it builds the week. Aldana cannot be scheduled to Park East and Westside in the same hours. The grid blocks the second save on its own. The multi-jobsite labor dashboard at the top of this page is what the Workforce Insights view looks like when multi-location is on. Every jobsite, every staffer, every hour, in one screen instead of four. The shared foreman pattern (Aldana split across Park East and Westside) is exactly what the system is built to handle. One staffer profile, schedule visibility across both sites, the AI Copilot checks for clashes at build time, and Park sees the split allocation in real time.

Walsh used to assign critical-path days from memory and seniority. Some of the crews he trusted were not actually showing up at the rate he thought they were. The Park East MEP sub had been missing 1 in 8 scheduled days — small enough nobody noticed week to week, big enough to slip a 4-week MEP rough-in by a full day or two over the run of the project. On the Reports & Analytics page, the reliability metric surfaces clock-ins divided by assignments across every staffer and every crew, trailing 90 days. Reliable crews get the critical-path Tuesday lifts and the Saturday pours. The lower-reliability crews get the non-critical-path days while you have the show-up conversation. The honest caveat we are required to put on the page: the metric counts a missed clock-in the same as a no-show. If a staffer worked the day but forgot to clock in on the app, the metric reads that as a miss. The fix is to drill into the underlying clock-entry log before drawing conclusions on a single crew. The metric is a screening tool — useful at the crew level, directional at the individual level. Combined with the show-up conversations, it makes promotion and critical-path assignment decisions data-grounded instead of memory-grounded.