Product Pricing About Blog Contact Request Access
← All posts

Why rostering is still broken in 2026

SchedulingWorkforce ManagementProduct

Ask any shift manager in hospitality, healthcare, retail, or logistics how they build their weekly schedule and you’ll hear a version of the same story: a copy-pasted spreadsheet, a handful of WhatsApp messages, a rostering tool that’s 80% there, and a lot of last-minute phone calls.

We’ve been talking to managers and operators for the better part of a year, and the pattern is remarkably consistent. The problem isn’t that rostering tools don’t exist — it’s that none of them respect the rules the way a good human scheduler would.

The three jobs a schedule has to do

A published roster has to simultaneously satisfy three very different stakeholders:

  1. The business — minimum coverage per position, skill mix, hours budget, compliance rules.
  2. The manager — fairness across the team, training opportunities, continuity on shifts.
  3. The employee — availability, preferred days off, hour targets, rest between shifts.

Every “simple” rostering tool pretends one of these layers is the only one that matters. Sling leans into employee self-service. Deputy leans into time tracking. Spreadsheets lean into manager control. None of them sit on top of all three and produce a schedule that works for everyone by default.

What actually changed

Two things changed in the last 18 months that make this problem newly solvable:

  • LLMs got good enough at constrained optimization problems. Not raw solving — we still lean on OR-Tools underneath — but at explaining the solution, spotting edge cases the constraints didn’t cover, and surfacing intelligent trade-offs to the manager.
  • Hourly teams got tired of piecing together four tools. They want one system of record, not another silo.

What CrewCentral is doing about it

Our bet is that the next generation of rostering isn’t a prettier UI on top of the same manual workflow. It’s a system where you define the shape of your week once — positions, templates, skills, coverage targets — and then generate, review, and publish a constraint-aware schedule in minutes, not hours.

And because so much of the world already lives in Sling, we sync bidirectionally, so you don’t have to rip out what’s working today to get the AI layer on top.

More on the constraint engine and how we think about explainability in upcoming posts. Until then, join the waitlist and we’ll get you access as soon as your region is ready.