Simply put, data-driven decision-making is integral to success in the modern era. Whether it’s availing of predictive analytics to optimize inventory based on historical data, tapping into AI and automation to recommend upsells, or growing the bottom line through insight monetization, there are countless examples of how data and analytics can improve business decisions. 

However, that doesn’t mean that human judgment and experience no longer have a role in decision-making. As machines become increasingly capable of making higher-level decisions, organizations and professionals alike will need to be more selective in determining which decisions can be addressed through data analytics and which should continue to be based on the human experience. 

This is one of the points made by Gartner’s Patrick Long in a recent eBook, “The Future of Decisions.” He stresses that companies must dissect and reengineer decisions to separate human input from machine involvement, enabling the two to work together when necessary. 

To help organizations wrap their heads around this, Long outlines three levels of decision-making engagement for humans versus machines:

Higher-Level Decision Support 

Individuals are responsible for decision-making at this level, drawing on experiences and bias, logic and reasoning, emotion, principles, and other uniquely human characteristics. While machines may be involved by providing visualizations, exploration, alerts, or additional support, the final decision is ultimately made by humans.

Augmented Machine Support

The augmentation at this level can take various forms. For example, perhaps the machine suggests, but the human decides. The inverse could also occur where the human suggests, but the machine ultimately decides, and it’s also possible that the human and machine could decide together. At this stage, the role of devices and AI is to provide the necessary recommendations and analytics that humans require to validate and explore potential outcomes.

Highly Automated Settings 

Machines can handle autonomous decision-making at this level, using tools such as predictions, forecasts, simulations, rules, optimization, or another form of AI. Because there may be scenarios in which guard rails are required, or humans may need to be involved, it’s essential that companies periodically check in with how machines are performing at this level.

You can read more on balancing machine capabilities and human expertise via this Forbes piece.