We ran it end to end on a representative demonstration portfolio: 18,000 finished-goods SKUs across 7 families, plus 1,201 raw materials. And no, you will not find a single "up to" on this page.
A quick word on how to read these. Each number carries the context that makes it honest: the portfolio it ran on, the period, the SKU count, and what it was measured against. That is deliberate. In a market full of “up to” claims, the figure that means something is the one you can trace, then reproduce on your own data.
Representative demonstration dataset: 18,000 SKUs, 7 families, 19,201 references total; not a client portfolio. Results vary by product mix, data quality and current policy, and are validated on your own data during the demonstration.
A distributor ran ERP min/max on thousands of SKUs and could not say why any number was set. We replaced it with deterministic policies and freed working capital while holding service.
A short-shelf-life portfolio was writing off stock every month. Pre-obsolescence detection caught the slow movers early and cut the write-downs.
Fill rate was stuck, and every attempt to lift it meant more inventory. We raised service across sites with no net inventory increase, by right-sizing each item.
It forecasts demand and sizes your inventory, sitting on top of your ERP. AI sharpens the forecast; the stock and cash decisions run deterministically, so the same inputs always give the same answer and you can explain any number months later.
Both, in the right places. Machine learning makes the forecast sharper, while the inventory and cash decisions stay deterministic. AI improves the number; it never decides it behind a curtain.
No. It sits above SAP, Oracle, Dynamics, NetSuite, Sage or your own system, optimizes monthly, and writes the settings back. Nothing to rip out, no rip-and-replace.
Anything that can export your demand and inventory history. We map it to your ERP once, and from then on it reads your data and writes its recommendations back where your planners already work.
Your sales or demand history, your current stock positions, and basic item and lead-time data. If your data is messy, that is normal; tidying it up is part of what we do, not a prerequisite you have to finish first.
The free 45-minute demo runs on your own data in a single sitting. A full on-premise install is measured in weeks, not the multi-quarter projects you may be used to, because there is nothing to rip out.
Nowhere. It runs on-premise, inside your walls, on your schedule. Your data never leaves, and every number stays reproducible and ready for audit.
We handle new and slow-moving items explicitly, leaning on the behavior of similar products and your planners' input rather than pretending the history exists. You still see how each of those numbers was built, exactly like the rest.
On a representative portfolio it reached 4.1% error across the whole book, with a median of 8.8% per SKU. We do not ask you to take that on faith; we rerun it on your data in the demo, so the accuracy you see is your own.
You see the opportunity on your own data inside the 45-minute demo. Most clients feel it in the stock and the cash within three to six months of going live.
Yes. It fits focused teams from about 5,000 SKUs up to large groups, across manufacturing, distribution, healthcare, energy and food.
Twenty years in the weeds, plus a method you can actually audit. We show the math instead of a black box, and we tie our engagement to results we agree up front; built by planners, for planners.
We run DemandIntel on your own data and walk through the forecast, the cash it would free and the risks it flags. It is a technical, strategic conversation, and nobody will push you to buy.
Each engagement combines a software licence with a short consulting phase to put it in place and train your team. We scope the figure to your size and situation after the demo, so you get a number tied to your own operation rather than a list price; the demo itself is free.
DemandIntel runs on-premise, inside your own environment, so your demand history, stock and costs never leave your walls; there is no cloud in the loop. Access follows your own controls, and in the free demo your export is used only to run that session and is not retained.
We support the deployment and your planners directly. New versions are installed on your schedule, in your environment, never pushed silently; and because every number is reproducible, an upgrade never quietly changes a figure you cannot trace.
What we build stays with you; the model and its settings live in your environment. If you continue, it is on a renewable licence, with your data and configuration yours to keep. No lock-in that holds your numbers hostage.
Book the free 45-minute demo and send us a data export when you are ready. That is the whole first step; everything else follows from what we find together.
Picture your next S&OP meeting: one number on screen that nobody argues with. See it free, on your own data, in 45 minutes; and if it does not hold up, we part as friends.