Recently, we have been seeing reports everywhere about how technologies such as Artificial Intelligence (AI) are dramatically changing the world. Unfortunately, what exactly is changing then often remains unclear.

One specific topic that Starcode is currently working on is the dynamic calculation and prediction of material requirements and capacity planning. In most ERP systems, MRP (Material Requirement Planning) is rather static. Here are some required data:

– The number of orders and quotes, or in other words what products are expected to be delivered and when.

– The bill of materials (BOM) of these products.

– The delivery time and/or production time for each part in the bill of materials.

– Production capacity.

These are just some of the many parameters you work with in today’s ERP world. For example, if we want to know what we need to purchase and produce in the next 3 months, we currently calculate this with MRP based on available orders and quotations.

In order to assemble products on time, you also need to have the required parts on time. In a stable world, delivery time should be consistent. However, events like Covid and subsequent supply chain problems have taught us that this is not always the case. Unfortunately, ERP systems, designed long ago, cannot take this into account.

In the system you have to specify the delivery time for each product, possibly by supplier, and this is then factored into the calculations by MRP. Suppose the standard delivery time is 20 days. Then for a sales order to be delivered on November 1, you need to consider the following schedule:

November 1 ready – assembly time (1 week) – delivery time (20 days) = final order date October 3.

We assume for a moment that there is no inventory and other reservations in this example.

Unfortunately, in practice things sometimes go wrong. Suppose we order in this example on October 2. Unfortunately, the delivery time of 20 days actually turns out to be 40 days due to current problems. Supplier: “Sorry, hopefully next time in 20 days”. This leads to postponement of production and delivery unless alternatives are available.

We have experienced such situations several times over the past 2 years. Even if suppliers indicate in advance that parts have longer delivery times, the current system cannot respond adequately. An example we experienced with a customer from the automotive industry:

Customers order a product with an expected delivery time of 12 weeks. This cannot be shifted around just like that, because otherwise it would lead to the postponement of an entire production line. But some parts were given a delivery time of at least 6 months. This caused major problems in the organization.

In order to still be able to deliver in such a situation, we actually want to purchase based on a ‘forecast’. Unfortunately, a standard ERP system does not provide a ‘forecast’ other than the orders and quotations entered. This is practically not a so-called forecast but hard data. The current ‘forecast’ often relies on sales team estimates and historical data in Excel.

Nevertheless, we still run the risk of buying the wrong parts if we don’t know which variant of a product the customer wants to order. This can lead to excess stock and the risk of unusable products (obsolete stock).

These situations encouraged Starcode to develop an ‘intelligent’ version of MRP using AI. Here we not only predict the ‘forecast’, but also calculate and forecast part delivery times and the impact on production capacity and orders.

Wondering how this works? Feel free to request a demo! In the coming weeks we will discuss this topic in more detail in other blogs.