The schedule is a forecast of the factory, and nobody scores it
Module · Production Planning and Scheduling
Every other forecast in the business is measured. The schedule is not, so the gap between the plan and the week is argued rather than attributed.
Prophesee Manufacturing scores the schedule and tests the sequence before it is released.
The factory is judged on adherence. Nobody benchmarks it.
Every other forecast in the business is measured against what happened. The schedule is not, and the sequence that drives most of the loss is still chosen by hand.
Sources: Godlan, OEE benchmark, data Jan to Dec 2025 (1,470+ discrete manufacturing operations); vendor published on its own installed base, so treat as indicative · Deloitte, Smart Manufacturing and Operations Survey, 2025 (n=600 US executives at companies above $500m revenue; fieldwork Aug to Sep 2024) · Qlector, How Leading Manufacturers Plan Production, Oct 2025 (largest Slovenian manufacturers, sample size not disclosed) · 3RDi review of published schedule adherence benchmarks, Aug 2026.
From publishing the schedule to scoring it
ForesightRun rates and changeover times were written at implementation. The line has been re-tooled twice since and the standard has not moved.
The measured distribution of run time and changeover by product, line and shift, set against the routing standard. Every plan is then built on what the line actually does.
PulseThe run produces hundreds of messages ranked by order number, so the scheduler works the top of the list and the line stops anyway.
Every works order tested continuously against material, capacity and labour. When the plan becomes unrunnable the alert names the binding constraint and the customer order behind it.
HorizonThe sequence is chosen by hand against due dates. The changeover cost of that choice is never priced against the batch size it saves.
Sequences generated and compared under mixed integer and constraint programming, priced on changeover, capacity and due date, with the projected curve shown before the orders are released.
NexusThe works order, the material, the quality event and the operator note live in four systems, so explaining a late order takes a morning.
Works orders, materials, downtime, quality events and shift notes resolved into one graph. Ask in plain language and get the constraint and the history behind it.
Turning production challenges into decisions
The sequence is chosen by hand against due dates, so changeover cost is never priced against the batch it saves.
HorizonThe run reschedules hundreds of works orders when nothing outside the factory has changed, and the floor stops trusting it.
Pulse HorizonAdherence is reported as a single number with no attribution, so the plan and the floor blame each other every week.
Foresight PulseHundreds of messages ranked by order number rather than by the customer order and revenue behind them.
PulseRun rates and changeover matrices were written at implementation and have not been measured against the line since.
Foresight Nexus14 AI applications that could be relevant
A sample of what becomes possible on the decision layer, not a fixed list: each application draws on the same data foundation and audit trail, and new ones are configured on the engines, not built from scratch.
The measured distribution of run time by product, line and shift, not the routing standard.
Expected yield and its spread by product and line, so the plan orders the right quantity.
Which works orders will miss, days before they do.
The gap between plan and week decomposed into plan error, material, capacity and labour.
The moment the plan becomes infeasible, with the binding constraint named.
Sequences that cost more in changeover than they save in batch size.
Ranked by the customer order behind the works order, routed to a named owner.
Orders rescheduled by the run without a corresponding change in the real world.
A sequence tested against changeover, capacity and due date before the works orders are released.
Model a shift pattern, a line addition or an outage and watch the plan update.
The date the plan reaches the adherence target, tracked against actuals.
Test lot size, time fence and horizon settings against plan stability.
Why this order is late, with the constraint and the history behind it.
Where the routing standard and the measured reality diverge, by operation.
A day when the schedule holds
Today: A sequence built by hand against due dates, with changeover absorbed silently.
ForesightMeasured cycle times replace the routing standard. Nine products are being planned twenty percent faster than they run.
Today: An exception list of four hundred messages, sorted by order number.
PulseA works order becomes unrunnable at 11:00. The alert names the material and the customer order behind it.
Today: An adherence number with no attribution, argued at the Monday meeting.
HorizonTwo sequences compared on changeover, capacity and due date. He releases the one that costs four hours less.
Today: A routing standard that has not been measured since the line was re-tooled.
ForesightThe week decomposed into plan error, material, capacity and labour. The Monday meeting takes ten minutes.
Tomorrow's sequence becomes today's decision.
Run the plan you published
We agree the metric and the baseline in week one, and measure the result on your data.