Understanding data vs. information, the role of human insight, and persistent knowledge in an AI-enabled world
As the Digital Transformation progresses and Artificial Intelligence (AI) gain traction in the corporate world, can we trust automated, AI-enabled automation to handle business decisions in highly complex supply chain environments?
In the report by VP Analyst, Tim Payne, Digital Business Requires Algorithmic Supply Chain Planning (November 2018), Gartner states the following regarding Algorithmic Supply Chain Planning:
“In this emerging digital era, the traditional, best-practice-focused, manually intensive, highly collaborative planning approach (analog planning) won’t extract value from the massive amount, granularity, and speed of data generated throughout an ever-extending enterprise business network.”
AI-enabled forecasting and planning automation allows businesses to:
- eliminate human error and bias
- achieve decision-making speed and scale
Which Decisions Should be Automated?
Planning, with the support of AI, requires you to understand where to draw the line between data and information. You should automate decision-making when backed by statistical models (data), but incorporate human insight (information) that algorithms cannot see. Forecasting and planning that rely on large amounts of fast-moving data should, naturally, be automated; while planning that engages information stored elsewhere (in the minds of leaders and planners), cannot.
Data vs. Information: Historical trends and patterns in demand can be found within historical data and are best detected by AI, while upcoming promotional plans and shifting customer/supplier relations are information human planners need to include.
Human Insight: Incorporating information (rather than data) into the plans is the key role of humans in the autonomous supply chain. Notions that exist in human brains should be incorporated into a planning system and treated as data.
Persistent Knowledge: The most advanced planning technology can retain new information that humans teach it, and treat that information as data. Using persistent knowledge, the system, once taught the information, can continuously improve performance. For example, if an upcoming promotion is planned, it is best to describe that promotion to the system rather than override the forecast. This allows the algorithm to optimize the plans going forward.
Exception Management: Even in autonomous supply chains, there are times when the reality of business and the market creates a situation that an AI-enabled tool deems infeasible. Exception management allows planners and leaders to look into those situations, evaluate business trade-offs, and decide on the best course of action.