The AI revolution in manufacturing requires human work

The AI revolution forms a paradigm shift for manufacturing. Most factory automation processes of the past were based on a rules-based approach, e.g. robotics programming with fixed set of scenarios and workflows which were top-down planned in ERP or MES systems. The downsides: very little flexibility in case of unexpected situations, huge efforts to prepare systems for high variance / low volume type of production.

Now comes AI

AI allows to reimagine manufacturing processes, across all functions of a factory. AI automates processes and machines, thus making production systems more flexible and adaptable. In a recent study, Boston Consulting Group has identified several example applications for the future usage of AI in manufacturing:


  • Machines self-optimize their parameters based on material input and process parameters
  • Automated guided vehicles carry parts, detect obstacles and adjust their routes
  • Warehouses will be automatically optimized based on material flow and inventory levels
  • Maintenance needs will be predicted by identifying failure patterns
  • Quality issues will be predicted by analyzing and learning from process data
  • Quality issues will be detected through image recognition.
  • Assistance systems suggest solutions to incidents based on earlier failure reports


By deploying the right combination of AI technologies, manufacturing companies can boost efficiency, increase flexibility, and improve processes towards self-optimizing operations. However, it is unlikely that AI systems will take over all jobs in manufacturing. If AI is deployed primarily to displace human workers, only short-term productivity gains will be realized.


Much needed: Human-Machine cooperation

In his recent book Human + Machine: Reimagining Work in the Age of AI, Accenture’s CTO Paul Daugherty highlights one thing: the most lasting, impactful performance boosts happen when human workers and AI-enabled smart machines work together. In this sense, AI systems are augmenting human capabilities and turn factories into smart factories:


  • Humans and smart machines collaborate closely.
  • Operational processes become more fluid and adaptive.
  • Human workers and machines are becoming partners, exploiting what each party does best.


But what exactly are the distinct tasks for Humans and Machines?

Humans are needed to develop, train and manage various AI applications. Machines, in turn, are extending human capabilities, for example to process and analyze large amounts of data in real time. The most important, and currently often neglected, part is what Daugherty calls “the missing middle”. Those activities where human-machine interaction takes place and where AI augmentation reshapes processes:


1. Amplification: the AI agent analyzes processes and provides data-driven insights, often using real-time data. Such insights will be presented through novel user interfaces such as smart glasses or smart watches. One example from manufacturing is to send maintenance alerts to service technicians:  the AI identifies failure patterns of a machine and sends maintenance alerts in realtime to the smartwatch of a service technician.


2. Interaction: the AI agent uses advanced interfaces such as voice-driven natural-language processing to drive interactions between or on behalf of people. Examples in manufacturing are personal assistance presenting step-by-step work instructions, adapted in real-time to fit to the context of an individual job, e.g. adapted to workers skill levels or presenting individualized product requirements. Thus, standard operating procedures (SOP) become Adaptive operating procedures (ADOP).

3. Embodiment: the AI agent is connected to smart machines. Lightweight robots with sensors, motors and actuators engage in physically collaborative work. Examples in manufacturing include package-carrying autonomous guided vehicles which automatically manage material flow inside a factory. The AGV closely collaborates with human workers and provides all needed material at the time when it is needed, e.g. in automotive assembly AGVs are a central part of flexible and modular production cells.


In the smart factory of the future, AI will automate processes and machines, empowering humans to respond to unplanned situations. Most critical situations will still require humans to decide. The AI will aggregate all historical machine and process data and propose actions. Humans will be able to make smart decisions based on the proposals of the AI. As a result, human workers and AI work alongside and bring various benefits to functions such as assembly, maintenance, quality, and logistics.