Family business Van Duuren not only has a good transport network, but also a good business network. In this way, the carrier knew that BPSOLUTIONS has a data science team that creates models. These models can predict developments based on Artificial Intelligence or AI (of which Machine Learning or ML is a part). An accurate look into the future helps organizations to reduce risks, improve services, seize new opportunities and increase revenue.
MODEL RECOGNIZES 90 PERCENT OF THE DELAYS
The data scientists of BPSOLUTIONS fed various Machine Learning models with all that information. Among other things, IBM Watson was used to assess the different models. That only happened after the data had been cleaned up. Trajectories like this usually involve some pollution. In this case, it turned out, for example, that there were different status numbers that had the same meaning. To get a good picture of which statuses really have an influence, they had to be added together. The best model was then able to correctly predict nine out of ten delays on the basis of test data. The machine can obtain insights from up to thirty variables super fast. The human brain is far from having that ability. The model turned out to achieve the best result when it is trained with data from the last year. If it now has to process new shipments, the result is again a correct prediction for 90 percent of the cases.
SHORTLIST TO TAKE ACTION EASIER AND FASTER
The big advantage for Van Duuren is that the new system now provides employees with a shortlist with potentially delayed shipments. So they no longer have to go through all orders. That saves a lot of time. But the most important result is that Van Duuren can now inform its customers in good time. This ensures greater understanding and higher customer satisfaction, which ultimately leads to more growth. Van Duuren can also use the list to be proactive in contact with partners: employees can ask why a shipment cannot be delivered on time. They can also obtain information about the new delivery time.
STEP-BY-STEP PLAN PROVIDES A PRACTICAL, PROFITABLE SOLUTION
To arrive at a well-functioning solution such as that at Van Duuren, BPSOLUTIONS takes roughly six steps:
- Having conversations with the customer: where are opportunities for improvement through the use of data and AI?
- Defining the business case.
- Collecting data.
- Cleaning and analyzing data.
- Developing a Machine Learning model.
- Implementation of the model with automatically generated predictions in an operational system, dashboard or other form, for example a report.
On average, approximately half a year is required for analysis, model development and implementation. This was also the case with Van Duuren.
The strength of BPSOLUTIONS is that the company is highly practical. The data science team is focused on developing solutions that deliver results. This approach helps customers improve processes and achieve growth. BPSOLUTIONS links expertise in the field of Artificial Intelligence to knowledge of technology. They have the people in-house who can realize the infrastructure necessary to get started and achieve goals.