Qu · MSME USE CASES
Six problem shapes we see repeatedly — combinatorial, constrained, and currently planned by hand.
SIX PROBLEM SHAPES
Each card lists the signals we look for. If two or three match your operation, it is worth a conversation.
The problem
Jobs compete for shared machines with sequence-dependent setups, due dates and shift constraints. Planners rebuild the schedule in a spreadsheet every morning.
Our approach
Model as a constrained assignment problem, solve with a hybrid annealer, and keep the planner in control of the final commit.
Signals this fits you
The problem
Multi-depot deliveries with time windows, vehicle capacities and driver hours. Route plans are built from habit and local knowledge.
Our approach
Formulate as capacitated routing with time windows, solve hybrid, and compare distance and vehicle count against the historical plan.
Signals this fits you
The problem
Reorder points set per SKU by rule of thumb, with lead-time variability and supplier minimums pulling in opposite directions.
Our approach
Joint optimisation across SKUs and locations under budget and service-level constraints, with demand uncertainty modelled explicitly.
Signals this fits you
The problem
Parts nested onto sheet, coil or fabric by an operator. Every percentage point of waste is material cost straight off the margin.
Our approach
Reformulate nesting as a packing problem with rotation and grain constraints, solved with quantum-inspired methods against the current layout.
Signals this fits you
The problem
Formulation and candidate screening runs on intuition plus slow classical simulation, so only a handful of options get tested.
Our approach
Variational simulation on small systems, used to triage a candidate list before classical methods and lab work take over.
Signals this fits you
The problem
Shipped devices use RSA or ECC with no inventory of where keys live, how they rotate, or whether firmware can be updated at all.
Our approach
Inventory cryptography across devices and firmware, rank by exposure and field lifetime, then test a PQC migration on one device class.
Signals this fits you
HOW A PILOT STARTS
Historical data in, the incumbent plan as the baseline, and identical time budgets for every solver. You see the comparison table before anything touches production — including the cases where the classical solver wins.
NEXT STEP
The readiness assessment takes two minutes and points at the most likely fit.