How Machine Learning is Transforming Supply Chain Management Before Our Very Own Eyes

Team Pecan

Want to learn more?

Contact us

Over the past years, our team here has encountered quite a few examples of how deep learning based predictive analytics was able to make a significant impact on business decisions of our customers. Quite a few of these examples had to do with Supply Chain Management, and how AI can really transform that industry.

Supply chain management (SCM), the management of the flow of goods and services, is critical in almost every industry today. AI and machine learning are defining the next generation of supply chain management. According to the report, State of Artificial Intelligence for Enterprises, Supply Chain and Operations was one of the top areas where businesses are driving revenue from AI investment. AI and Machine learning algorithms, and the models they are based on, excel at finding anomalies, patterns and predictive insights in large amounts of data. Many supply chain challenges are time, cost and resource constraint-based, making machine learning an ideal technology to solve them. For example, Amazon’s Kiva robotics relying on machine learning to improve accuracy, speed and scale and DHL relying on AI and machine learning to power their Predictive Network Management system that identifies the top factors influencing shipment delays.

Gartner predicts that by 2020, 95% of Supply Chain Planning (SCP) vendors will be relying on supervised and unsupervised machine learning in their solutions. Gartner is also predicting by 2023 intelligent algorithms, and AI techniques will be an embedded or augmented component across 25% of all supply chain technology solutions.

The application of AI into Supply Chain related-tasks holds high potential for boosting top-line and bottom-line value. Valuable time and money are wasted on trivial supply chain related-tasks that are conducted operationally by humans. Businesses spend on doing manual, paper-based processes and checks, chasing invoice exceptions, discrepancies and errors and responding to supplier inquiries. This loss- can be thousands of hours, during the work year, that businesses are throwing away by processing papers, fixing purchase orders and replying to suppliers. ( 2017)

There are many ways how AI and ML can be applied within SCM activities. Including Chatbots for Operational Procurement, ML for Supply Chain Planning (SCP), for Warehouse Management,  Autonomous Vehicles for Logistics and Shipping, ML and Predictive Analytics for Supplier Selection and Supplier Relationship Management (SRM) and for sales and demand forecasting: 

Machine Learning (ML) for Supply Chain Planning (SCP):

Supply chain planning is a crucial activity within SCM strategy. Having intelligent work tools for building concrete plans is a must in today’s business world. Applying ML within SCP can assist with forecasting within inventory, demand and supply. Correctly applying through SCM work tools, can revolutionize the agility and optimization of supply chain decision-making. Intelligent algorithms and machine-to-machine analysis of big data sets can optimize the delivery of goods while balancing supply and demand, which wouldn’t require human analysis, but rather action setting for parameters of success.

Machine Learning for Warehouse Management: 

SCP success is greatly affected by proper warehouse and inventory-based management. Regardless of demand forecasting, supply flaws: overstocking or understocking, can be a disaster for just about any consumer-based company/retailer.  ML provides an endless loop of forecasting, which bears a constantly self-improving output.

ML and Predictive Analytics for Supplier Selection and Supplier Relationship Management (SRM):

Supplier selection and sourcing from the right suppliers is an increasing concern for enhancing supply chain sustainability and vitally important. Supplier-related risks have become the ball and chain for globally visible brands. Machine Learning and intelligible algorithms can help to predict the best possible scenario for supplier selection and risk management and during every single supplier interaction. Data sets, generated from SRM actions, such as supplier assessments, audits, and credit scoring provide an important basis for further decisions regarding a supplier. With the help of Machine Learning, supplier selection can be more predictive and intelligible; creating a platform for success from the very first collaborations. All of this information can be easily available for human inspections but generated through machine-to-machine automation; providing multiple ‘best supplier scenarios’ based on whatever parameters, in which, the user prefers.

ML and Predictive Analytics for sales and demand forecasting:

Brick-and-mortar grocery stores are always in a delicate dance with purchasing and sales forecasting. Predict a little over, and grocers are stuck with overstocked, perishable goods. Predict a little under, and popular items quickly sell out, leaving money on the table and unpleased customers. The problem becomes even more complex as retailers add new locations with unique needs, new products, ever transitioning seasonal tastes, and unpredictable product marketing.  For decades, retailers have extrapolated demand by looking at historical sales data—an obviously imperfect methodology that skews demand forecasts downward, since it doesn’t measure unmet demand. Advanced machine-learning algorithms overcome this problem. The algorithms build demand probability curves using sales and inventory data, making cost-benefit calculations that evaluate the risk of waste against the risk of out-of-stocks. 

Unlike standard supply-chain software systems, machine-learning solutions can collect, analyze, and adjust large data sets from a wide range of sources. Machine learning algorithms can make demand forecasts based not just on historical sales data but also on other influencing parameters: 

  • Internal factors: such as advertising campaigns and store-opening times.
  • External factors: such as local weather and public holidays.

The calculations are done at a much more granular level than standard systems are able to do which brings many benefits including that retailers can determine the effect of each parameter on each stock keeping unit (SKU) in each store (and in each distribution center, where relevant) on a daily basis.

In Pecan, our customers can create predictive models using machine learning within days and with no need to data scientists or programming skills. Pecan allows them to apply ML within their SCM initiatives. Moreover, now companies can answer questions such as; which factors most influence the prediction and the likelihood prediction per individuals. For example, for supplier selection problem, why supplier X predicted that way? what are the factors which influenced the most?; For sale demand problem, what product is most likely to buy next week? 

Right before we end this article, we’d like to note the next frontier which we are tackling here at Pecan.

Quite a few of our customers are using AI to answer “What if” questions. We call it Predictive Hypothesis. The Pecan Predictive Hypotheses engine gives you the power to use AI to scientifically predict the outcomes of various business decisions, and how they will affect each and every aspect of your bottom line. An example this “What If” question would be: What will be the impact on the sales demand when raising 5% of the price of product X? 

We will tell you more about it on our next blog post series.


Medium, 6 Applications of Artificial Intelligence for your Supply Chain. October 2017. By Kodiak Rating Community.
Forbes, How To Improve Supply Chains With Machine Learning: 10 Proven Ways, April 2019. By Louis Columbus.

It's time to plug your organization into its future.

Try Pecan