Artificial Intelligence (AI) and machine learning have gained significant momentum over the past few years, and with businesses from all industries continuing to embrace the technology at rapid rates, adoption stands to increase even further. In fact, Gartner predicts that by 2020, 85 percent of all customer interactions will be managed without a human. Furthermore, Narrative Science found that 80 percent of executives believe AI improves worker performance and creates jobs.
Supply Chain Benefits
For manufacturers, AI and machine learning are particularly important technologies, because they have the power to disrupt the very way products are made, moved and sold. Rather than guessing about which materials are most appropriate for various products, mistakenly selecting a subpar logistics provider or spending months researching which markets are most applicable for certain items, AI and machine learning-powered solutions have the ability to continually and accurately recognize trends and make data-driven decisions, all without any human intervention.
By incorporating the automation and efficiencies of AI and machine learning into their operations, manufacturers have an opportunity to disrupt their entire supply chain. They can produce products at lower costs and avoid processing issues such as delayed deliveries. Additionally, they can reduce their inventory levels by having access to more intelligent planning systems that dynamically update according to warehouse status and/or current market needs.
The Role of Process Mining
Data analytics has allowed some manufacturers to make more informed decisions through utilizing massive amounts of data collected to uncover hidden patterns, correlations and customer preferences. However, it has limitations. While analytics software can shed some insight on business operations, it has always required a hypothesis on where to shine the light in order to see where things work well or work poorly. Users need to pose the questions first to the analytics software. Without those questions, analytics cannot uncover anything. Also, global supply chains and manufacturing operations tend to get so complex that it is hard to understand what is really going on in detail.
And that is where process mining, fueled by AI and machine learning, comes in. A new type of Big Data Analytics, process mining leverages the digital footprint organizations leave behind in their IT systems to automatically reengineer any organization’s supply chain processes. This provides manufacturers with complete transparency into how processes are working in real life, enabling them to pinpoint business process inefficiencies. This way you see the “as-is” state, including all the analytics. Process mining can explain why processes are broken and how to fix them by giving corporate data a full body scan and unbiased analytics to solve problems manufacturers did not even know they had.
Eliminating Major Inefficiencies
Given the various departments, individuals and moving parts that make up a manufacturing organization, business process oversights can occur daily, and sometimes those oversights remain undetected for months on end. These cross-departmental inefficiencies can seriously harm manufacturing operations by slowing organizational throughput, impacting working capital, extending customer response time and creating team performance bottlenecks.
Consider the following scenario: Most supply chain organizations feel they have a good handle on their business processes, however in reality, processes such as purchase-to-pay may not be working as designed. Perhaps purchase orders with incorrect data are being received by vendors 20 percent of the time, for instance which leads to manual rework. This is a serious issue. Another example are planned delivery lead times. They have to be maintained on an individual material code basis for every location so the production planning knows which time to expect to be replenish every individual material that goes into production. Therefore, their accuracy is absolutely crucial for a just in time supply chain. In case the planned time lead time for any material is too short, this leads to very costly out of stock situations; in case it is too long it drives up inventory. Working with numerous large manufacturers, we found out that this is an area with tremendous potential for optimization. Leveraging process mining and machine learning, it is possible to very precisely predict delivery lead times, thus making the supply chain more robust and efficient.
These are just two examples of a major inefficiency that manufacturers can quickly and easily identify with AI and machine learning-powered solutions before it becomes a dire problem. Automation capabilities mean teams of workers won’t have to spend twenty minutes manually correcting each and every purchase order that was sent with incorrect information and planning is not based on feelings.
Embracing Automation and Transparency
As AI and machine learning continue to have a significant impact on the way manufacturers execute their business processes, some worry that the automated technology might render human workers obsolete. However that’s not where the industry is headed. Instead, AI and machine learning offer an invaluable ability to assist workers by taking on the responsibilities too tedious and time-consuming for humans. As a result, manufacturers and their workers can focus their time and intelligence on more worthy tasks, and leverage that increased productivity to better compete against international organizations.
For manufacturers looking to optimize their operations and spur greater innovation by implementing AI and machine learning technologies, a crucial first step is achieving total transparency. Process Mining is a helpful tool to leverage here, as it takes the digital footprint manufacturers already have in their IT systems to provide complete visibility into how processes are actually working. Perhaps it’s unclear which suppliers are being most supportive, for example, or why one sales office or plant is consistently outperforming others. Process Mining can answer these questions, identifying which teams have the best outcomes, and where inefficiencies are leading to bottlenecks and wasted time.
Once manufacturers achieve complete operational transparency, they can start to gradually automate various business process steps by implementing AI and machine learning-powered solutions. In doing so, they stand to realize a significant margin uplift due to reduced inventory levels, lower manufacturing costs and faster throughput times. Best of all, manufacturers will be able to drive greater customer satisfaction and loyalty levels due to fewer mistakes and delays.
Alex Rinke is co-founder and CEO of Celonis.
