Recently I was interviewed by the Intelligent Enterprise Leaders Alliance for their latest publication. Industry experts explore the current state of AI adoption across the supply chain in the latest publication, “AI & Analytics in Supply Chain,” addressing misconceptions and highlighting impactful applications. These experts include Lora Cecere, Cathy Roberson, Omid Ghamami, Sheri Hinish, and associations including the Reverse Logistics Association, Institute for Supply Management (ISM), CSCMP, IWLA, and more. The full transcript of my interview is reproduced below. You can get a copy of the full publication at the Intelligent Enterprise Leaders Alliance website. 

=====

Key Considerations for Leveraging AI and Predictive Analytics to Cut Through the Complexity of Supply Chains

Interview with

Lisa Anderson
Founder & President, LMA Consulting Group Inc.

How can businesses better align sales, operations, and supply chain teams around real-time inventory visibility?

Businesses can better align sales, operations, and supply chain teams around real-time inventory visibility with a SIOP (Sales Inventory Operations Planning) process and supporting systems and technologies. The purpose of SIOP is to align cross-functional teams (sales, operations, supply chain, finance, R&D, etc.) on a collaborative forecast, long-term production and rough-cut capacity plans, and resulting inventory plan. To gain real-time visibility, the key is to deploy the appropriate ERP system functionality and supporting technologies. Inventory transactions will occur in the ERP system to ship, receive, move, transfer and adjust inventory. Without advanced technology, users must perform transactions in real-time; however, if they roll out barcoding, RFID, EDI, blockchain and/or related technologies, they automate the process of achieving real-time transactions. On the other hand, deploying these technologies will produce inaccurate real-time information if users are not accompanied with process disciplines. For example, you must perform steps of the process in the correct order. One resource cannot miss barcoding an item or the next user will have to resolve an error.

What areas of the supply chain are most ripe for improvement through predictive or prescriptive analytics?

Techniques are those requiring heavy data analysis, optimization, and trend spotting. Sales forecasting is ideal for predictive capabilities as it is predicting demand. Also related to demand, predictive analytics will analyze sales patterns, highlight exceptions, and point out opportunities. Similarly, advanced planning analysis predicts supply chain requirements. For example, transportation and route planning predicts the best route, mode of transportation, and carrier for a specific route based on multiple variables including driver availability, cost, service etc. In global logistics, an advanced planning system can predict the next disruption and re-route or change the mode of transportation or sourcing location to proactively address the issue and optimize cost and service. In manufacturing, an advanced planning system can predict and reallocate production to optimize service, cost, capabilities, and manufacturing performance.

Data governance and integrity principles are important. In complex and custom environments, rules and parameters must be incorporated into the process to fuel machine learning to ensure directionally correct results. Assuming these political issues are addressed, machine learning will pick up on trends and patterns that would be missed otherwise.

What’s your view on the role of machine learning in demand forecasting and its ability to improve accuracy over traditional methods?

Machine learning can provide added value to demand forecasting as it is best equipped for this type of function. Although the capability is superior to traditional methods for spotting demand patterns, it does not always produce a better forecast because it must be accompanied with a strong process and data protocols. Generally, using machine learning will produce a better base forecast. If the company uses a SIOP (Sales Inventory Operations Planning) process to minimally a demand plan review process to address exceptions, review trends, and incorporate sales feedback and customer forecasts, machine learning will perform best.

What’s your perspective on AI’s ability to manage complexity in supply chains, particularly for global and multi-tier supply networks?

Artificial intelligence will be of paramount importance in cutting through the complexity of supply chains and providing results for review. AI can crunch through more data than the best supply chain resource and review multiple scenarios on a rapid basis to optimize inventory, service, cost, and responsiveness. The key to success will be training and educating high-skilled supply chain resources to know how to feed AI the appropriate information and review and adjust the results to maximize supply chain performance. You cannot assume that AI will provide “best fit” recommendations every time and incorporate everything your top planner(s) will look for and adjust. On the other hand, it will provide a solid base far quicker than your best planner due to its capabilities and capacity to evaluate alternatives and scenarios. The best companies will marry up advanced supply chain systems with AI capabilities with top notch planners to maximize results and mitigate risks. The trick will be how to develop these high-skill planners over the long term if basic planning functions can be taken over by AI. Proactive companies will develop and train planning professionals to build these high-skilled planners and leaders.

Reproduced in full from the interview in ‘AI & Analytics in Supply Chain’, Intelligent Enterprise Leaders Alliance