Edge Digitalisation by AI in Supply Chain and Logistics
Published by : Industrial Automation
Jasbir Singh elaborates on how Artificial Intelligence can help address the increasing complexities in supply chain.
Artificial intelligence (AI) started entry into research level work to develop and produce ‘intelligent machines’ that are capable to learn, memorise, analyse by processing data, and act independently replacing human intelligence. Artificial intelligence is becoming more powerful in every process, and continuing to advance. Technology trains the machines to behave like humans perform their functions, or sometime exceed in capabilities.
The potential of AI is enhancing supply chain and logistic activities and helping develop strategies to streamline operations. AI in such cases replaces intelligent human actions or acts of behaviour, for problem-solving by continuous learning from routine operations. AI in digitalised machine functions having learned the action-based capabilities, which make it capable to display intelligence while performing the required actions in real time delivery. The major factors in supply chain management include forecasting, sales and demand, spend analytics and logistics network optimisation. A lot of time was being spent by executives to generate reports on regular basis/periodically, analyse data to decrease cost and increase business revenues. This attracted companies to use artificial intelligence to act or prompt leaders by information based actions instantly. AI has produced great output in improving human decision-making processes by enhancing productivity in various business endeavours. It improves ability to recognise business patterns, learn business phenomena, collect information, and analyse data intelligently.
There are two predominant ways to apply AI in Supply Chain and Logistics:
1. Autonomous: Fully automated supply chain and logistic processes that can operate without human intervention, i.e., Robotic operation for all key processes (critical and non- critical functions); and
2. Assistance and collaboration: Interdependent processes supported with human intelligence decision-making for data analysis and virtual assistance in operations to minimise error, i.e., Cobotic assistance to support human during operation. Here the output is mainly controlled by humans.
Managing the supply chain
These are two major categories of AI capabilities and all logistic processes are falling within every business processes in current and future implementation. Target is for enhancing productivity in operation in all unit operations. By application of AI in operations, companies move from manual to fully automated functions. It is well said that any operation which can be performed manually is possible to be fully automated by the application of AI and machine learning processes. This is true with companies having material movement within and external supply chain and logistics related tasks as part of their system. The application of AI into any Supply Chain related task creates high potential value for boosting tangible returns.
Dependency on conventional system to collect data for generating various reports consumes great deal of time by people in supply chain department which always have possibility of errors. This adds hefty delay to improve the core processes that need to act faster to react with market demand and improvise to deliver. In the present era of globalisation large companies operate with multiple number of sourcing, production and distribution to further large numbers both within internally and with partners or directly to consigners worldwide. Supply chains management is much different now than a few years ago, and it will continue to evolve in present intense competitive economy of world. The innovation in supply chain management uses the concept of new logistic pop-up warehouses, ship-from-store models, innovations at speed and scale, and brings major change in global supply chain management (SCM). Supply chain management for company demands to effectively manage the growing product portfolios that require using many thousands of stock-keeping unit number (SKUs). The number of SKU continues to rise as companies endeavour to meet customer expectations for multiple variations in product configurations, range of chemicals and pharmaceutical API/ingredients.
Organisations typically take undesirable steps like:
1. Storage costs increase because companies have to pay for longer-than-necessary leases to accommodate the sales volume
2. Companies have to rack up additional shipping costs to shuffle balance excess inventory at other warehouse locations, and
3. Companies can do nothing, resulting in over-loaded/underutilised space and inventor.
Pop-up warehousing effectively addresses the most of these common scenarios. It addresses the variable capacity issues that arise during common peak storage and distribution scenarios. Best suited AI in supply chain management essentially requires scalable solution for any enterprise using cognitive automation, which is a subset of artificial intelligence that uses advanced technologies like natural language processing, emotion recognition, data mining, and cognitive reasoning to emulate human intelligence.
Cognitive automation using AI
Cognitive automation platform is essentially a scalable AI that can process terabytes or even petabytes of data. This platform also can execute thousands of data wriggle across any number of internal or external systems, then aggregate and normalise that data; in other words called as a cognitive data layer. AI with ML algorithms are then applied to produce desired recommendations for optimal actions to improve supply chain speed and cost-efficiency. Cognitive automation simulates slow, difficult and unusable data in lightning speed, which was traditionally done by humans, often in Excel spread sheets.
The right example of implementation of AI in supply chain management is by considering cognitive automation in pharmaceutical and medical device company that uses the process/technology for ‘available-to-promise (ATP)’ metrics. ATP improves the forecast delivery date close to or more than 90%. This in turn improves end-to-end visibility of global demand and supply of product by collecting and processing real time data across multiple systems and provides dynamic display at dash board for managers. It helps to predict lead times through machine learning which is required for accurate ATP dates and quantities, especially important for repeat orders. AI enhances supply chain to monitor near real-time changes of supply and demand that could impact ATP dates and quantity, and also recommend corrective actions to prevent impacts on ATP dates and quantities.
Product identification using barcode and numbering system
Cognitive automation platform delivers data analysis down to the SKU level. It is a unique string of letters and numbers representing each product in a warehouse/seller’s inventory. Standard practice for SKU creation often has chosen approach that makes things the least complex for the attendant at warehouse/seller’s delivery personal who need to interpret the SKU’s data for details of product; whereas Universal Product Code (UPC) is another identifying coding system associated with product. The major difference between UPCs and SKUs, where SKUs are seller-dependent, UPCs remain the same even when the respective products are being delivered to consumers. These numbers facilitates smooth and fastest movement of product from one place to other helps in supply chain management.
Without real-time information, AI tool shall be making bad decisions faster, given that:
1. Stale data results in poor decision-making.
2. AI needs access to external and downstream data to improve better result than that of a traditional system.
3. Support for high consumer service levels at lowest possible cost.
4. AI tool for decision making by considering change vs the cost of change.
5. Decision making process must be continuous by self-learning and self-monitoring.
6. AI system interacts with problem continuously and fine tune corrective action as required.
7. AI engines must use autonomous decision-making engines.
8. Facility for users to monitor and even override AI decisions when necessary.
9. Use of cognitive automation to process huge volumes of data very quickly and must be able to make smart decisions, fast, and on a massive scale.
Important factors of AI in supply chain management and logistics
Improve end-to-end visibility: where cognitive automation platform addresses to deal with huge data enter across applications to create a single, virtualised data layer. This layer explains cause and effect, bottlenecks and opportunities for further improvement. It operates with real-time data, rather than stale information that may be days, weeks or even months old. Actionable output: Despite investing high cost in collecting huge data and business intelligence solutions, many companies fail to derive result required to make right decisions in a timely manner at greater speed and agility.
Less or no dependency on manual/human work: Traditionally supply chain professionals spend large hours for gathering data from multiple systems and utilising business intelligence tools or spreadsheets to figure out action plan. To avoid this difficult task to handle too much data, too many applications and too many variables to account for a cognitive automation quickly gives actionable result with detail analysis.
Improved decision-making: Cognitive automation automates and augments on decisions with AI-driven predictions and recommendations suggested for optimal actions to improve performance. It disseminates result for implications across various attributes such as time, cost and revenue. Cognitive automation platform can be improved to act autonomously and also gather learning over time to continuously improve on recommendations as any conditions change.
Machine learning techniques are the baseline to streamline the whole logistics control in supply chain management using AI to automate processes such as load forecasting and vehicle scheduling. AI software includes functionality that is segment of computer software decides the way to collect real-time information from the raw data, which in turn delivers key decisions.
Increasing complexities in supply chain require fundamentally learning and action-based capabilities of autonomous AI systems to achieve the speed and agility in today's markets.
Jasbir Singh is an Automation Expert with experience in Factory Automation and Line Automation in a large production house. He is an Implementation Strategist, Business Coach and a regular writer on automation, AI, robotics, digital technology, network communication, IIoT, wireless communication, blockchain and use of advance digital technology. Jasbir has a long association with industry to improve factory automation in production lines for productivity improvement in India and overseas by advising and also transforming into digital platform by use of AI. He may be reached by mail at: email@example.com