Everyone’s talking about AI. But are we making the most of the knowledge we already have?

Everyone’s talking about AI. But are we making the most of the knowledge we already have?

German industry continues to struggle to stay competitive. Although there are early signs of an economic recovery, cost pressures remain high. At the same time, companies are investing in digitalisation, automation and artificial intelligence in the hope of making better decisions, improving efficiency and lowering costs.

These investments are necessary. But sometimes it’s worth asking a different question. Are we already making use of the information in our own systems?

An example from inventory management shows why this question matters.

A manufacturer updates one of its products. It has new technical features, a new item number and perhaps revised packaging. The replacement item is set up in the ERP system without any consumption history. The planning team has to work out how much of it will be needed.

So the team estimates initial stock levels, sets safety stock levels and places orders. Meanwhile, substantial quantities of the original item may still be sitting in the warehouse.

What’s surprising is that the demand itself is often the same. Only the item number has changed.

The original item may have been on sale for years. Its consumption history shows seasonal fluctuations, the quantities needed and how demand changes over time. Yet companies don’t always use that information systematically when switching from one item to another.

We’re discussing AI-powered forecasting while leaving data we already have unused elsewhere.

 

Demand history from suitable predecessor items can be carried over to their replacements. This requires taking into account any changes in product features, lead times or sales conditions, as well as remaining stock and open orders.

 

With suitable planning methods and empirical simulation, companies can use this data to develop better-informed forecasts and set safety stock levels. DISKOVER, for example, uses item lineage records to help transfer historical demand data to replacement items.

 

What may seem like a minor detail in inventory planning can have a significant financial impact. Excess stock ties up capital, and every shortage puts the ability to fulfil orders at risk.

 

Especially in challenging economic times, companies should ask questions that go beyond which new technologies they need.

They should also ask how much untapped potential there is in the data they already have.

Sometimes, smarter planning can start simply by making systematic use of what we already know.

FAQ – Frequently Asked Questions

How can you forecast demand for a replacement item with no consumption history of its own?

The demand history of a suitable predecessor item can serve as the basis for the initial forecast. It shows how much was typically needed, how much demand fluctuated and whether there were any seasonal patterns. Once consumption data for the replacement item becomes available, you can check how well your assumptions match actual demand.

A predecessor item’s demand history is particularly useful when a product has been technically updated but its applications and customer base remain similar. If product features or sales conditions change significantly, check which demand volumes and patterns still apply to the replacement item.

Which inventory planning parameters should be reassessed when switching to a replacement item?

Review replenishment lead times, lot sizes and service level targets for the replacement item. These factors help determine how much stock to hold at launch and how much safety stock is needed to achieve the target service level.

The forecast estimates expected demand, while safety stock covers the remaining uncertainty. Optimising both together helps identify the combination that achieves the required service level with the lowest possible inventory.

How can you avoid holding unnecessary stock of both items during a product changeover?

Plan the phase-out of the predecessor item and the phase-in of its replacement together. Define the overlap period, plan how demand will gradually shift to the replacement item and account for remaining stock and open orders. Using up remaining stock of the predecessor can reduce the need to purchase or produce the replacement item, provided the predecessor meets the technical requirements and customers still accept it as a substitute.

DISKOVER uses item lineage records to link the demand histories of multiple predecessor items to a single replacement item. The planning process can also account for existing stock and open planning items for the predecessors, such as sales orders, reservations, purchase orders and production orders. This helps determine how much of the replacement item still needs to be purchased or produced.

Picture of Dirk Ungerechts

Dirk Ungerechts