Agriculture is the least digitized of all major industries, according to the McKinsey Global Institute’s Digitization Index. But a recent study by the same firm entitled “Human + Machine: A new era of automation in manufacturing,” looked at agriculture’s potential for automation and the results suggest that this has little to do with the fundamental tasks and activities that make up farming. The report set out to not only evaluate the automation potential of major global industries but also to analyze the forces that determine how and why automation is or is not implemented.
In a comprehensive assessment of the potential for automation, agriculture ranked fourth with a 57% potential for automation, behind accommodation and food services (73%), manufacturing (60%), and transportation and warehousing (60%).
Agriculture beat retail trade (53%), mining (51%), construction (47%), and finance and insurance (43%), among others.
The scores were developed by studying manufacturing work in 46 countries and evaluating the frequency of certain activities and the ability of these tasks to be automated, creating an aggregate score as a percentage. The tasks are management, expertise, human interface, unpredictable physical labor, data collection, data processing, and predictable physical labor.
For agriculture, the most prominent of these activities is “unpredictable physical” activity, which the study rates at approximately 40% automation potential.
Under today's unique circumstances, AgFunder is re-opening Fund III for a limited time to enable investors to join our mission and invest alongside us as LPs in a second close. Learn more here.
Notably, the ratings were also based on existing technology, with no speculation about potential technological advancements coming in the future, meaning that 57% of agricultural activities are automatable today.
Where in the Supply Chain?
The work that the study deems most readily automatable is the physical work of processing crops and data collection. Notably, data processing is assessed to have less potential for automation than data collection.
The activities listed in the study from the most potential for automation to least potential for agriculture are:
- Data collection
- Predictable physical labor
- Data processing
- Unpredictable physical labor
These tasks take the concept of automation far beyond the common framework of field robotics. Looking at agrifood tech through this lens, the majority of the technologies covered by the term contain some move toward automation.
Accommodation and food services ranked first of all the industries included in the study, ostensibly because the most prominent activity type was deemed “predictable physical” activity which has 100% automation potential.
Indeed cooking robots and restaurants with no human cooks like Zume Pizza are coming online and raising large rounds.
Where in the World?
Also notable in the study: McKinsey does not find much variation in automation potential based on geography.
McKinsey reports that “81% of the world’s automatable manufacturing hours and 49 percent of automatable labor value reside in developing countries,” meaning the potential impact of automation in these countries is immense. According to the report, 62% of this highly automatable labor is in India and China.
McKinsey predicts that conversion to automation will be rapid and disruptive in developing countries like India and China, but not until the cost of automation dips low enough to compete with the low cost of labor.
Where to Start?
The study points out that the quality of the work is a major factor in deciding what to automate, beyond simple cost-benefit analysis or technical feasibility. Even when automation is a cheaper option, if the work output is of lesser quality or slower speed, automation may not be a good choice.
The authors explain that in order to access what kind of automation would best suit a business, it is important to know from which stage of maturity the business is starting along what it calls the “spectrum of automation maturity.”
Businesses with low automation maturity may benefit from starting to build automated systems from scratch, while those with high automation maturity are more likely to need to work on gleaning the most benefit out of automation technology already present.
Want to read more about automation in Agrifood tech?