Is AI needed to improve SCM performance?

In a recent article “Shouldn’t GenAI Mean New Gen?” Supply Chain Shaman, Lora Cecere wrote:

“My belief is that GenAI on the Graph, through the deployment of outside-in thinking, offers the opportunity to reduce the number of planners by 80-85%........Why do we need hundreds and thousands of planners?”

Lora is quite right on the opportunity to reduce Planner numbers but perhaps wrong on the reasons. A hypothetical perfectly flowing supply chain is one in which the materials move exactly in line with customer demand providing 100% service levels with no static stock. The barrier to that is not lack of AI but inaccurate forecast-push MRP and the accompanying backorder averting schedule interventions, unreliable processing activities and the reasonable need, in many industries, for batch size quantities above one unit - all of which cause materials to stop moving.

Minimising batch sizes and maximising the reliability of processing activities is what much of Lean is all about (eg. SMED, TPM, TQM, Standard Work, Poke-Yoke etc etc) and getting materials to move in line with demand is delivered by Enterprise-wide Pull - often known as Demand Driven MRP. It is because Enterprise-wide Pull moves materials in line with demand (instead of an inaccurate forecast) that there is no avalanche of MRP exception messages and the consequent need for schedule interventions/expediting – thereby also significantly reducing Planner workload.

Later in her article Lora writes:

“The supply chain planner has the lowest satisfaction level of any supply chain employee…”

which isn’t very surprising given that they spend most of their time amending schedules to compensate for the inaccuracy of the forecasts that were used to create the master production schedules. Although Planner workload is reduced by Enterprise-wide Pull the content of the role becomes far more interesting, value-add and satisfying. Instead of all that expediting the Planner is now able to focus on tuning (or conditioning) the flow of materials in line with demand (through parameter and exceptional event management), capacity planning/S&OP, product life-cycle management and perhaps, if working in a demand-driven extended supply chain, collaborating with partners at customers and suppliers to enhance end-to-end flow.

In contrast to another of Lora’s quotes from the article

“I find the industry awash with opinions but short on research-backed answers.....”

the evidence for the effectiveness of Enterprise-wide Pull (Demand Driven MRP) is multiple successful implementations, see DDI Case Studies and numerous papers from academic research institutions such as

Demand Driven MRP: assessment of a new approach to materials management

Material Management without Forecasting: From MRP to Demand Driven MRP

Investigation of potential added value of DDMRP in planning under uncertainty at finite capacity

The reason why Enterprise-wide Pull / Demand Driven MRP is so effective is because it mimics Flow - it moves materials in response to demand just like water moves in response to gravity and slope, see Neither water or supply chains need "Big Tech" to tell them how to flow.

AI will undoubtedly find a role in supply chain management but its not needed for improving forecast accuracy, achieving your planned service levels with 40% less inventory, elimination of expediting and the need for catch-up capacity. That just needs Demand Driven Institute - compliant software operating through your ERP.