Autonomous supply-chain systems — also known as self-driving supply chains — make most decisions automatically, without human intervention. There is much foundational work for end users to do, with increasing benefits at each stage. Meanwhile, solution providers are continually building out various elements of autonomy across ever more functions, which are becoming more deeply embedded and intelligent.
Automation of individual tasks involves steps on the path to fuller autonomy. We’re already seeing some fully autonomous planning and limited forms of autonomous execution. The types of decisions being automated are diverse, including:
- design (generative design and preference prediction);
- sourcing and procurement (cognitive sourcing and procurement automation);
- automated “touchless” demand planning and inventory optimization;
- merchandizing and category management (agile category resets, promotion planning, automated planogram optimization and compliance checking);
- logistics (precise ETA, disruption prediction and corrective actions);
- warehousing;and
- service (predictive maintenance driving automated parts ordering and service scheduling).
Beyond individual functions, solution providers are striving toward more end-to-end autonomous planning and execution, ultimately including multi-tier capabilities. It will take a while to fully realize that kind of ambitious vision.
Meanwhile, the predecessor of autonomous supply chain execution can be found in control towers, which provide supply chain-wide visibility into the status of orders and shipments and alerts in the event of delays. Solution providers have been extending control-tower capabilities in several dimensions, representing these seven steps toward autonomy: 
- Inventory tracking. Near-real-time visibility into inventory levels at plants, distribution centers and retail across the chain.
- Manufacturing status. Visibility into actual production status at suppliers and plants, including delays or changes.
- Predictive analytics. Predicting delays in production and logistics using data and signals such as materials shortages, equipment breakdowns, weather, traffic and rerouting of vehicles and vessels.
- Impact analysis and prioritization.Understanding the impact of various delays and prioritizing the ones that need attention.
- Resolution suggestions. Some systems are starting to suggest potential resolutions and show side by side the expected impact of each possible resolution. For example, expediting might enable earlier production, but at a higher logistics cost. This will eventually go beyond simple tradeoffs. Using artificial intelligence and machine learning along with chain-wide data and awareness, systems will be able to find more creative solutions that humans are able to discover, thereby reducing or eliminating traditional tradeoffs.
- Automated execution. Execute recommended actions with the click of a button. Behind the scenes, relevant planning and execution systems are automatically updated.
- Full automation.With machine learning, the system gets smarter over time. With increasing confidence that the system is providing the right answers, time after time, users will eventually be more willing to hand over the reins. This won’t happen all at once, but rather within constrained boundaries at first, addressing discrete issues such as orders below a specific size and minimum confidence level. We expect those boundaries will be tunable by end users, based on their confidence in the system and appetite for risk.
The Role of Digitization and Data Quality Initiatives in Getting to Autonomy
Data quality and a digital supply chain are prerequisites for companies moving to autonomous supply-chain capabilities. It can be difficult to make the transition from paper-based systems, such as e-mail and spreadsheets, directly to autonomous systems. Transactions must be digitized and reliably captured by systems, with consistent data and syntax, before autonomous supply chain capabilities are possible. Lack of sufficient data quality and completeness is an inhibitor to autonomy for most companies. Investments in digitization and data quality will pay off in multiple ways, well before full autonomy is reached.
For a deep dive on the evolution towards autonomous systems, in the context of managing demand-supply balancing, see our Agile Demand-Supply Alignment series.
To view other articles from this issue of the brief, click here.