In 1998, when I leveraged advanced algorithms to procure MRO products for the DND IT infrastructure, I learned something early: Algorithms are an extension of human experience and expertise multiplied.
For example – and this is the crucial difference between using an agent-based versus equation-based model, I knew only by experience that the time of day an order was placed for a part significantly impacted cost and SLA performance.
The challenge to that point was that field technician systems and incentives were based on the number of calls they did daily versus when they ordered the replacement part. As a result, they would sandbag all of their part orders till the end of the day to maximize their call response efficiency.
By the time the buyers received the orders, the price for a $100 product could be as high as $1,000, and because it was being sourced south of the border, next-day delivery would not happen, resulting in an SLA performance of 50% versus the contracted 90%.
It also meant that the procurement department’s investment in technology was of little value because said efficiencies could only start when the order was received. Not even today’s more advanced AI tech could effectively address this issue.
So, what is our next step?
Procurement Insights 
What will prevent AI from successfully transforming procurement efficiency and effectiveness?
Posted on September 2, 2023
0
In 1998, when I leveraged advanced algorithms to procure MRO products for the DND IT infrastructure, I learned something early: Algorithms are an extension of human experience and expertise multiplied.
For example – and this is the crucial difference between using an agent-based versus equation-based model, I knew only by experience that the time of day an order was placed for a part significantly impacted cost and SLA performance.
The challenge to that point was that field technician systems and incentives were based on the number of calls they did daily versus when they ordered the replacement part. As a result, they would sandbag all of their part orders till the end of the day to maximize their call response efficiency.
By the time the buyers received the orders, the price for a $100 product could be as high as $1,000, and because it was being sourced south of the border, next-day delivery would not happen, resulting in an SLA performance of 50% versus the contracted 90%.
It also meant that the procurement department’s investment in technology was of little value because said efficiencies could only start when the order was received. Not even today’s more advanced AI tech could effectively address this issue.
So, what is our next step?
Share this:
Related