In the age of advanced analytics and machine learning, humans can be in the loop, on the loop or out of the loop.
In supply chain, when humans are “in the loop,” they are the primary decision makers. Based on understanding — sometimes limited — of current complex conditions in the supply chain, they arrive at what they perceive as the best decision.
When the humans are “on the loop,” they rely heavily on the power of advanced analytics but continue to be the primary decision maker and final arbitrator of actions.
In the “out of the loop” model, advanced analytics and machine learning take over full decision making and execution. Humans take a supervisory role, only intervening when needed. Humans can feasibly be out of the loop in numerous supply chain processes, such as forecasting, inventory positioning, available to promise, dynamic routing, factory scheduling and many more. Deploying this model has been a driving force for massive investments in advanced analytics and machine learning platforms, and data science talent.
But despite the drive toward automation, when it comes to decision making, there’s always going to be a need for a mixture of “in the loop,” “on the loop” and “out of the loop” models. Different business requirements and priorities, different time horizons and different constraints will determine the exact level of contribution of human or machine intelligence in making the decisions.
While companies embrace decision-making automation, they must simultaneously devise strategies to capture human knowledge. The need to capture human knowledge is even more critical, as supply chains cope with talent retirement and increased mobility, and increasing emphasis on technical skills, sometimes at the expense of domain knowledge.
A supply chain organization will gain a competitive edge if it adopts a framework that mutually augments human and machine intelligence. When humans teach, machines can learn. When machines recommend, humans can supervise. When humans react, machines can refine. When machines falter, humans can take over. All with the goal of improving collective decision making.
In a recent report, Gartner introduced the reciprocal human-machine augmentation framework. The framework is defined as:
Continuous sharing of knowledge between humans and machines that increases their collective ability to make better decisions.
In the report, we introduce three strategies that support reciprocal augmentation:
The three strategies are (see Figure 1):
- Crowdsourcing: With this strategy, companies can extend the principles of crowdsourcing to include humans and machines as equal contributors to the wisdom of the crowd. Humans can bring in their insights based on their domain knowledge. Conversely, machines can become an active member of the crowd, contributing analytics-generated insights. The human-machine combined insights can then be used to improve decision quality.
- Process Mining: Leveraging process mining, machines will augment humans’ decision making by identifying automation opportunities and aligning operational decisions with supply chain strategies. Conversely, through process mining, organizations can capture human domain knowledge by tracking potentially value-adding deviations from defined processes.
- Data Literacy: By embracing data literacy, organizations can teach their staff to understand and speak data and analytics. By “speaking” data and analytics, humans can augment machines with better articulation of business requirements and with supervision of machine performance. By understanding data, humans augment their decisions with analytics insights and recommendations.
As we observe the solemn two-year mark since the onset of the COVID-19 pandemic, supply chain organizations more than ever are keen on adopting decision models that can be equally effective in managing daily conditions and responding to major disruptions. The most effective and adaptive decision models are those that leverage the combined power of human and machine intelligence. The reciprocal human-machine augmentation framework creates a virtuous circle that ensures that the sum of the two is bigger than their individual contribution.
We explore improving decision quality with reciprocal human-machine augmentation in the November executive report (available to Gartner clients) and in the accompanying podcast, which is available on Spotify, Apple Podcasts and Google Podcasts.
Vice President, Distinguished Analyst
Gartner Supply Chain
The Gartner Blog Network provides an opportunity for Gartner analysts to test ideas and move research forward. Because the content posted by Gartner analysts on this site does not undergo our standard editorial review, all comments or opinions expressed hereunder are those of the individual contributors and do not represent the views of Gartner, Inc. or its management.