Manufacturers should leverage AI to take advantage of the current uncertainty and improve performance in 2021.

By John Sturdivant, Director, AI Success, DataRobot

Companies in the manufacturing industry have made something extremely clear: the challenge of making things for a profit became harder than ever in 2020. Given the volatility in demand, disruptions in operations, and generally depressed economic activity, the manufacturing winners will be those who can run the leanest operation and maximize return on assets.

As such, 2021 won’t be a year of expansion and innovation, but rather, a year of discipline and protecting profitability. Focus on getting the most out of existing capital will be paramount and AI will play a big part for those that do this well. In 2021, the most successful manufacturers will use AI to forecast customer demand better, manage supply chains more efficiently, and maximize output from installed equipment — these are all necessities for manufacturing excellence and why the following themes will be the three themes of AI in manufacturing in 2021.

Theme 1: Predicting Demand Amidst Heightened Complexity

As the pandemic continues into 2021, and as consumer tastes, needs, and resources systemically shift in the post-pandemic world, predicting who will buy what and when will become a significant challenge for manufacturers aiming to satisfy demand efficiently.

The COVID-19 pandemic changed many trends and patterns in demand forecasting data. 2021 forecasts will have black swan events in their lookback windows (e.g., March 2020), they will be affected by products still-recovering from demand shocks (e.g., aerospace), and correlations that once had signal will no longer be as useful in predicting the future (e.g. the flu season and demand for syringes or the school year and demand for classroom supplies).

However, getting better answers in demand forecasting is more important than ever. Manufacturers recovering from a difficult 2020 will have smaller reserves of working capital to manage, so producing exactly what customers want, when they want it will be critical to strong financial performance. Producing too much or at the wrong time will limit flexibility and tie up much-needed cash in inventory while producing too little will squander revenue opportunities and extend the path to recovery.

Demand forecasting with time series AI models that can be redesigned and retrained rapidly will help the manufacturing CFO build a more resilient, responsive, and insightful demand forecasting process and better manage production plans for a more profitable 2021.

Theme 2: Sophisticated Supply Chain Management

In the same interest of managing working capital better while meeting customer demand, sophisticated supply chain management practices will allow the COO to ensure production can meet all targets and supply chain disruptions won’t derail operations.

For example, raw materials consumption in the manufacturing of complex products will be an area of opportunity with AI production planning tools. Such tools will be able to forecast the consumption, breakage, and shrinkage of these items with high precision and facilitate better procurement and delivery planning – conserving cash and ensuring stock of materials.

Procurement functions will also be able to use AI models to make better decisions and better manage vendors. Repetitive tasks like invoice processing or QA/QC can be automated with AI-powered tools; predictive tasks like anticipating on-time performance and inventory availability can be accomplished with the analytical advantages of machine learning; and AI-powered text analysis can analyze vendor communications for potential red flags such as delivery disruptions.

Theme 3: AI-Powered OEE Maximization Techniques

In an effort to maximize return on assets during a period of reduced expansion and upgrade capex, advanced Overall Equipment Effectiveness (OEE) techniques that fully leverage AI will deliver superior performance to the operators willing to embrace new methods. Many engineering and management decisions made in the manufacturing environment over-index on judgment and intuition of a few experienced leaders, and don’t adequately consider the truths and trends hidden in the data. These decisions range from reliability and maintenance planning, to production staffing, to production yield optimization.

For example, shutdown and offline maintenance planning can benefit significantly from AI-powered risk models. More informed decisions about reliability risk for every system can lead to better scoping decisions for offline maintenance windows, which will create more availability, better reliability, and higher OEE. System maintenance histories, instrumentation data, and production data can be used to yield highly accurate predictive models for this type of predictive maintenance application. In this example, system availability will see an immediate boost, improving return on assets starting with the first impacted offline maintenance window.

Manufacturing Excellence with AI in 2021

In 2021, the manufacturers who are embracing AI as a way to improve overall performance and financial returns will be steps ahead of the ones that aren’t. Many operational problems created by the uncertainty and rapid change of the pandemic can be mitigated or solved with AI tools such as operational excellence, supply chain management, and production planning. However, courage is required as the answers can change existing ways of working dramatically.

John Sturdivant Datarobot, Industry Today
John Sturdivant

John Sturdivant, Director, AI Success, DataRobot

Previous articleJanuary 21, 2021
Next articleGlobal File Systems Come to the Manufacturing Sector