Imagine you can control your production lines with a mouse click or a touchscreen swipe to optimize your supply chain in real-time – along with the ability to enjoy a seamless, customized experience. 

With next-generation technologies disrupting our lives day by day, we are all trying to distinguish whether machine learning, artificial intelligence, automated decision making, etc. are just buzzwords or practical capabilities that can be integrated into our daily lives.

  • Will machines do our jobs on our behalf just like how they always seem to do in Hollywood movies? 
  • Is it really possible to fully automate decision making using ML and AI algorithms? 
  • How can we start making use of fast-developing technologies like ML and AI algorithms in our daily jobs today?

As supply chain professionals, in this rapidly changing world, we deal with many analytical challenges when it comes to planning and execution: 

  • uncertainty regarding future demand
  • bias in every aspect of the supply chain 
  • mismatched and missing data between different sources
  • lack of visibility and control over the complete network

Gartner identifies the 4 Evils of Supply Chain Planning (uncertainty, bias, data, model) in their article “Digital Supply Chain Planning: The Art of the Possible Should Focus on Reducing the Four Evils of Planning”

How to tackle the 4 Evils with the help of AI and ML algorithms:

All of these are issues getting in the way of making fast and effective decisions about how much inventory to hold, how much to purchase when, or how to best use the production capacities to maximize sales and profits while meeting customer expectations. 

Machine Learning (ML) and Artificial Intelligence (AI) are here to help companies make better decisions. 

83% of early AI adopters have already achieved substantial (30%) or moderate (53%) economic benefits – Deloitte

72% of business leaders say AI can enable humans to concentrate on more meaningful work – PwC


In today’s world, nobody can avoid or ignore uncertainty. What is important in an era dominated by machine learning and artificial intelligence is knowing how to handle the impact of uncertainty with our newly found tools in addition to the traditionally proven approaches.

  • Always on-demand sensing is one factor to be boosted with the use of AI. In addition to traditionally used historical data, it is now possible to include impacts of the hourly or daily changes in weather, promotions, store footfall, customer inventory, and point-of-sale data. Early exception detection and predictive alerts can be provided using always-fresh data received via automated data integration from both enterprise and external sources.
  • AI and ML algorithms are also utilized for updating forecasts for the remaining period based on short term trends and to identify gaps in sales targets. The products with the highest excess stock are selected and optimum sales price is recommended based on shipment history and price elasticity analysis. With the aid of connected data sources, profitability impact can also be reported.
  • Pricing decisions are essential for revenue, profit, and inventory projections. Being able to review possible short and long-term impacts of pricing scenarios can be eye-opening during hectic times. Advanced algorithms provide optimum price and activity recommendations which is a significant help to any planner.

Uncertainty is not only a significant factor in demand forecasting but also a part of supply planning. AI and ML algorithms help include lead time variability impact in safety stock & reorder point calculations. The external effects on lead time variability such as seasonality, weather, and road conditions, order frequency, or order amount can be considered as well in supply plans with the use of machine learning.


Humans’ decisions are inherently biased. To eliminate the impact of human-induced bias on planning, we can make use of technology.

 For instance, to analyze the performance of a promotion, we need to understand the following: 

  • sudden drops and increases in shipment history, 
  • stock-out impact on out-of-stock (OOS) SKU after stock replenishment, 
  • amplification impact on other similar products during the OOS period of that product… 

All these can be identified using AI and ML algorithms without being subject to errors caused by human bias!

  • Automated or user-selected supply plan scenarios optimized for multiple objectives like best service, minimized cost, maximized revenue, etc. help us in our efforts to make unbiased decisions.

Planners still may need to edit plans manually, however, these edits can be collected to learn from and leveraged to continuously improve the model to a point where manual edits are no longer needed. 

  • Multi-functional planning capabilities bring multi-functional teams together on a single collaboration platform so that teams can make their decisions using data-driven insights and prescriptive analytics to reduce bias while building a corporate memory across the company.


Manually maintained data in disparate systems result in discrepancies in master and transactional data. With the use of AI and ML algorithms, master data discrepancies and missing or inconsistent historical data can be identified automatically.

  • While we don’t recommend automating the master data correction process fully, a corrective action list can be provided. Sales loss due to stock-outs can be calculated with AI algorithms and the lacking shipment history for stock-out periods can be auto-filled based on previous shipment data
  • Orders left open for ages, inconsistent MOQ/IOQ, decimal typos… 

They all happen at times and can be easily detected and corrected automatically with range & reasonableness checks.

The best way to eliminate data discrepancies is to use real-time data. With flexible and automated data integration capabilities, it is easy to keep the ever-updated data always fresh and clean.


Even with the best-in-class optimization capabilities, if the model is not a good representation of the physical world, the planning output won’t be realistic and useful. Constructing a Digital Twin of your supply chain connecting end-to-end from vendors to customers in addition to the solutions listed above is beneficial in this case.

  •  A scalable digital platform on the web and a comprehensive data model for the entire value chain combined with a data exchange mechanism with multiple transaction systems provides unmatched visibility and understanding of the total network.
  • Horizontal alignment across demand, inventory, supply, procurement, manufacturing, fulfillment, and transportation and vertical alignment across strategic, tactical, and operational decisions can be maintained powered by fact-based and unbiased insights.

Although AI & ML might still seem to be only alive in science fiction movies, it is here and it is time to use it to tackle the evils in supply chain planning. 

By 2020insights-driven businesses will steal $1.2 trillion per annum from their less-informed peers – Forrester

At Solvoyo, we are aware of these challenges faced by Supply Chain professionals and we have the technology and experience to address these! 

We constantly update and transform our AI&ML technology in our forecasting and optimization engines to continuously improve our customers’ supply chain planning capabilities. 

Book a consultation session to plan your supply chain transformation journey.

Gartner “Digital Supply Chain Planning: The Art of the Possible Should Focus on Reducing the Four Evils of Planning” Tim Payne, 30 September 2019

Gartner does not endorse any vendor, product, or service depicted in its research publications, and does not advise technology users to select only those vendors with the highest ratings or other designation. Gartner research publications consist of the opinions of Gartner’s research organization and should not be construed as statements of fact. Gartner disclaims all warranties, expressed or implied, with respect to this research, including any warranties of merchantability or fitness for a particular purpose.