Check out the on-demand sessions from the Low-Code/No-Code Summit to learn how to successfully innovate and achieve efficiency by upskilling and scaling citizen developers. Watch now.
The modern global supply chain is defined by scale — billions of transactions and terabytes of data across multiple systems, with businesses generating more every moment. Traditional supply chain management (SCM) practices are quickly being outmatched by the ceaseless onslaught of information.
When a problem arises with inventory costs or availability, financial and demand planners dive into Excel or legacy SCM tools in an attempt to pinpoint the issues. This approach is like looking for the proverbial needle in a haystack. The sheer volume, velocity, and variety of data defy human efforts to understand the dynamics and right the ship.
AI has emerged as a hot topic in supply chain management to handle precisely this challenge. Innovative organizations are applying artificial intelligence and machine learning against vast sets of supply chain data to unearth insights into problems and performance that are effectively beyond the reach of even the most skilled planning professionals.
AI holds tremendous promise to optimize these processes. In fact, Gartner has found that 25 percent of organizations had begun AI initiatives through 2017, up from 10 percent two years earlier. Firms in pharmaceuticals, consumer packaged goods, manufacturing, and other industries are looking to move beyond relatively simplistic SCM tools built on static business rules that inhibit the ability to optimize and scale.
Intelligent Security Summit
Learn the critical role of AI & ML in cybersecurity and industry specific case studies on December 8. Register for your free pass today.
A common question I hear is, “How do we get started?” I’d like to offer three suggestions.
Target a specific problem
For a first project, it’s best to identify a specific supply chain issue that could be solved with AI. That helps focus efforts and resources on a single problem, rather than throwing spaghetti at the wall. Naturally, you’ll want to select a significant pain point with implications for your supply chain efficiency, customer satisfaction, and bottom line.
For instance, let’s say a global CPG company has challenges meeting service level agreements with its retailer customers. The company can face stiff penalties under its SLAs if stock is not delivered on time and in full. Applying AI to that specific issue has the CPG company on the fast track to resolving its service level fulfillment issues.
You may have a dozen potential projects for AI across your supply chain, from planning to production, packaging, warehousing, distribution, and logistics. Targeting one in particular positions you for the best results while minimizing the risk that ill-defined experimentations end up on the back burner. By selecting a discrete project, you can build on initial successes and learnings to apply AI in other areas.
Build AI readiness
Data is a critical ingredient of AI readiness. Ideally, companies apply AI to large data sets of deep granularity — for instance, three years of data with granularity down to the daily transactional level by SKU locations, orders, plants, raw materials, customers, and more.
Because you may draw data from different systems — ERP, MRP, CRM, and others — you ideally have time frame consistency across those data sets. In other words, if you have six months of raw materials data, one year of inventory data, and three years of sales data, AI can become problematic.
It’s a good idea to ensure your data collection and storage mechanisms are geared for highly granular data. Even if you’re not ready for AI today, you’ll be prepared down the road as AI becomes a competitive differentiator.
I’m often asked whether an organization needs a data science team in place to get started with AI. It can certainly help, but it’s not a prerequisite at the start. What you do need is people with source system and domain expertise to help map out the AI landscape.
Consider whether technology partners can help
For companies that decide supply chain AI is right for them (which can depend on scale, complexity, etc.), success depends heavily on the technology partners they choose and the services those providers offer. Because AI is still an emerging innovation, it’s essential to ensure the technology partner you choose is equipped to work with you through obstacles and devise solutions that address your needs.
As with any enterprise software selection, you should do due diligence and find a partner who actively addresses your needs. For instance, you may find the AI software is not able to handle the necessary data aggregation and validation. Others may have limitations in their ability to collect data, which is best accomplished by crawling source systems hundreds of times a day. The depth of analytics and underlying technologies are other aspects to consider.
An AI partnership is a two-way street. The likelihood of success increases exponentially if your organization has a visionary CIO and other leaders who boldly embrace innovation. Prepare to be surprised, as AI can unearth data-driven insights that obliterate assumptions you might have taken as gospel.
It’s important to recognize that supply chain AI is something of a journey into uncharted territory. But it’s without a doubt the most exciting innovation the supply chain has seen in decades. While not every company will need it, those for whom supply chain AI can make a difference should consider getting started sooner rather than later. Companies that have taken the leap are already seeing huge dividends as AI reshapes their supply chain operations.
Kaushal Dave is the vice president of cognitive (AI) supply chain solutions at Aera Technology, a cognitive technology for the self-driving enterprise.
VentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.