If you run an EMS operation of any size, you’ve seen the scheduling dance. Someone — usually a supervisor or ops manager — sits down with a spreadsheet, a seniority list, a stack of time-off requests, and a mental model of who can work with whom, who’s about to hit overtime, and which shifts nobody wants.
It works. Until it doesn’t.
The problem isn’t that this person is bad at their job. The problem is that the job is computationally hard. In operations research, workforce scheduling with constraints is a well-studied class of problem — and it’s NP-hard in the general case. That’s a formal way of saying: as the number of employees, shifts, constraints, and preferences grows, the number of possible schedules explodes. A human can find a feasible schedule. Finding an optimal one — minimizing cost, maximizing fairness, satisfying all constraints — is beyond what intuition can do at scale.
This isn’t a theoretical concern. When your 40-person agency has 8 unit types, 3 shift lengths, seniority-based bidding, mandatory rest periods, certification requirements, and a union contract with 14 relevant clauses, the number of valid schedule permutations is astronomical. The person building your schedule isn’t optimizing — they’re satisficing. They’re finding the first thing that works and moving on, because they have a hundred other things to do today.
The result is predictable: uneven shift distribution, preventable overtime, grievances about fairness, and a persistent sense that the schedule is something that happens to people rather than for them.
Constraint-based optimization solvers — the same kind of math used in airline crew scheduling, logistics routing, and manufacturing planning — can do this work. You define the hard constraints (coverage requirements, rest periods, certifications). You define the soft objectives (minimize overtime, honor seniority, distribute undesirable shifts fairly). The solver searches the solution space and returns schedules that satisfy the constraints and optimize the objectives.
This isn’t AI hype. It’s well-established applied mathematics that other industries adopted decades ago. EMS just hasn’t had access to it in a form that fits operationally and financially.
That’s what we’re building at Shiftbid.