Integrating artificial intelligence to support lean principles is a lever for both transforming manufacturing processes and improving efficiency and quality, however it requires a structured approach that combines technological innovation and methodological rigour.

What is artificial intelligence: its foundations and evolution

The birth of artificial intelligence dates back to the 1950s, when a number of pioneers began wondering whether a machine could actually “think”. One of the first was Alan Turing, with his famous test, while in 1956, at the Dartmouth conference, John McCarthy coined the term “artificial intelligence”.

The first systems were based on very rigid logical rules: perfect for mathematics, yet not really useful for dealing with more complex situations. In the 1980s, new impetus came from artificial neural networks, inspired by the functioning of the human brain and capable of simulating more flexible learning processes.

In the early 2000s, with the boom in big data and the increase in computing power, machine learning and, subsequently, deep learning began to be developed: techniques that made it possible for machines to learn from data and evolve over time.

We are now in the era of generative AI, which is not limited to analysing massive amounts of data, but rather is capable of generating different types of content.

Lean and AI: practical approaches and technology

The aim of lean philosophy is to reduce waste, value people, and achieve continuous improvement. Artificial intelligence, on the other hand, offers the possibility to process enormous amounts of data in real time, recognising patterns or signals that may otherwise not be picked up by humans. Integrating these two approaches offers the opportunity to develop more efficient and flexible production systems.

A number of studies have highlighted the potential benefits of integrating AI and lean:

• Optimised resource management: AI facilitates the effective allocation of production resources, quickly adapting capacity and materials based on varying demand and operational priorities. This improves flexibility and reduces waste and downtime.
• Predictive maintenance: by analysing data from sensors and machinery, AI can anticipate failures or anomalies before they cause machine downtime. This means targeted service can be planned, reducing downtime and extending the working life of the equipment.
• Faster and more accurate decisions: by generating insights based on company data, AI enhances decision-making, helping identify critical issues, manage actions, and optimise resources effectively. 
   

Artificial intelligence makes data a strategic asset for continuous improvement, helping identify the key to reducing waste, delays, and inefficiencies in business processes.

Risks and threats of adopting AI

While offering numerous advantages, adopting AI in lean processes does involve certain critical issues that, if not managed carefully, can jeopardise the success of the project:

• Data quality: AI models are extremely sensitive to the quality of the data they are trained on. Noisy, incomplete, or inconsistent data can lead to incorrect predictions or the inability to detect anomalies.
• Cultural resistance: in many manufacturing environments, AI is seen as a “black box” or a threat. This can lead to mistrust and slow down adoption, especially if there is no active staff training and engagement.
• Self-referential models: AI learns from historical data and therefore can easily replicate biases, inefficiencies, or mistakes from the past. If left unchecked, these mechanisms reduce AI’s ability to drive meaningful continuous improvement.
• Inconsistent outputs (“hallucinations”): some models, particularly generative AI, can produce results that are incorrect or out-of-context. To avoid making the wrong decisions, it is essential to always maintain human control and adopt rigorous validation.
• Technological dependence: the excessive use of AI can reduce people’s autonomy, delegating too much power to technology. The risk is that the flexibility and problem-solving ability underlying lean thinking will be lost.

An effective roadmap for sustainably integrating AI and lean

Integrating artificial intelligence into lean processes requires a gradual and structured approach centred around people, data, and technology. Effective implementation can be divided into several phases:

1. Clearly map processes: a thorough understanding of activities is needed to identify where AI can generate real value, avoiding superficial or misaligned actions.
2. Ensure data quality and availability: it is essential to collect as much data as possible. Data cleansing is equally important, though, as without a reliable database, even the most advanced algorithm will fail.
3. Involve people from the outset: change cannot be imposed from above. Training and feedback are needed so that AI can be perceived as a support tool and not a threat.
4. Initiate pilot schemes: start with limited but meaningful use cases, measuring the benefits and building internal know-how around the use of the technology, so as to gain trust in the technology.
5. Integrate AI into the lean PDCA cycle: the introduction of AI also needs to follow the iterative logic of the Plan-Do-Check-Act cycle, adapting to the results and feedback it receives as inputs.
6. Update models: AI is not static, and the models need to be updated over time to maintain expected performance.
7. Promote human-machine collaboration: the goal is not to replace human thought, but rather to support it. AI supports the ability to observe, analyse, and improve, aspects that are typical of a lean mindset.

Conclusions

The integration of artificial intelligence into lean contexts is a real opportunity to improve the efficiency, quality, and responsiveness of manufacturing processes. However, technology on its own is not enough: AI is a powerful resource that requires a conscious and structured approach, becoming a tool that serves processes and people, and not vice versa.

References:

¹. https://retrocausal.ai/blog/how-ai-is-shaping-the-future-of-lean-manufacturing/
². https://amfg.ai/2024/01/22/unlocking-the-synergy-between-ai-and-lean-manufacturing/