Supply as a signal, not a spreadsheet

Manual steps remain between consumption at a location and an order to a supplier. EIS already stores stock documents, requests, balances and write-offs, but the presence of a record does not guarantee a complete or timely signal.

A spreadsheet is a week late

The classic scheme: a shift notices something is short, the manager adds it to a request, the request goes to purchasing, purchasing merges it with the rest. By the time a decision is made the original observation is stale, and the reason it ran out has been lost.

The problem is not people’s discipline. The problem is that data travels through memory and retelling instead of being recorded at the moment of the event.

What counts as a signal

For future forecasting, a signal has to be a structured event: item, location, time, quantity and context. Not “we ran out of milk”, but a confirmed consumption event or stock correction with a known source.

• Sale, receipt, transfer, write-off and correction need a date, location and item.
• A deviation can only be calculated after those events are checked for completeness.
• Manual balance changes must remain visible or the forecast will mistake a correction for real demand.

Forecasting does not remove the human

Demand forecasting is the obvious application of the AI layer, which is exactly why it is easy to overrate. We work from the assumption that automation prepares a decision and a person confirms it: purchasing stays an area of responsibility, not an autopilot.

What breaks today

The honest picture at the review date: some data arrives through integrations, some is entered manually in EIS, and reference items still need matching between systems. Forecasting is therefore a next stage; a person confirms purchasing decisions.