Predictive Quality Control: A Strategic Ally for Operational Excellence through the Lean Method and Artificial Intelligence

In the contemporary industrial landscape, the pursuit of operational excellence through The Lean method is a constant. Companies strive to optimize processes, reduce waste, and, above all, ensure the highest quality of their products and services. In this context, the Predictive Quality Control (PQC) emerging as one of the most promising frontiers, representing a fundamental evolution from traditional quality control approaches. It is no longer just about identifying downstream defects, but about anticipating and preventing them, a concept that is the basis of Operating philosophy Lean Thinking.

Understanding quality with the Lean method: not just the absence of defects 

Before discovering predictive quality control, it is essential to understand the lean method, an operational philosophy born at Toyota that revolutionized the way production and services are thought of. The lean method is based on the systematic elimination of waste (muda), the creation of value for the customer, and continuous improvement (kaizen). This approach aims to optimize workflows, reduce lead times and minimize non-value-adding activities, always keeping the end customer's needs at the center.

In this context, it's essential to take a step back and reflect on what we mean by “quality.” The quality of a product or service is defined by its ability to meet Critical-to-Quality (CTQ) characteristics from the Customer's perspective. This means going beyond mere compliance with technical specifications, embracing the entire Customer Experience.

In this perspective, a quality system is the collection of all resources (people, tools, approaches, etc.) that enable an organization to produce quality products or services, thereby guaranteeing customer satisfaction with a view to continuous improvement. To build a robust quality system, it is necessary to define an organizational structure that allows, on one hand, to pursue and continuously improve customer satisfaction and, on the other hand, to make operational methods increasingly effective and efficient.

But what is meant by Quality? 

Let’s look at it from different perspectives: 

• Transcendental: understood as beauty or a philosophical concept. 

• Based on the productmeasured by the nameplate value of a quality driver, such as a PC's HD capability or the speed of an internet connection. 

• Based on usage: fitness for purpose, such as bicycle tires depending on usage or the portion size of fresh ready-made meals. 

• Based on productionconformance, for example the size of a hole or the fill level of a bottle. 

• Based on the valueA comparison between utility and price, like the service of a phone operator or a generic drug. 

The path to creating quality products or services unfolds through four macro-phases: 

1. Quality PlanningAnalyze the market and clearly identify our Target Customers with the goal of planning products that create demand, according to predefined quality targets and specifications. 

1. Designing QualityCreate design processes focused on building quality through prototypes to anticipate potential design problems and develop countermeasures. 

1. Manufacture QualityControl and improve production stages to build quality through the process. 

1. Sales and Quality Services: check quality through tests and surveys, sell the product, establish a post-sales support/service system, handle repairs and complaints quickly and reliably. 

The Hidden Cost of Poor Quality: The Impact of CoPQ on Company Revenue

The economic and reputational impact of a defect increases as you get closer to the end customer, a fundamental principle of the lean method which emphasizes prevention at the source. Identifying a defect during the production process has much less severe consequences compared to when the end customer discovers it.

 This concept is at the core of Cost of Poor Quality (CoPQ). These costs, which are often hidden and difficult to measure (intangible), include lost sales, late deliveries, reduced customer loyalty, inventory, excessive lead times, and costly design changes. Traditional quality costs, on the other hand, are more tangible and easier to measure, such as quality control, warranty repairs, scrap, and rework. This data highlights a massive waste of resources but also a potential that, if saved, becomes available to companies to increase their competitiveness in the market. Based on our project experience, we estimate that this amounts to 10–15% of companies’ revenue. 

From Traditional Quality Control to Predictive Quality Control 

The Quality Control (QC) traditional is focused on improving the ability of the production process to not generate defects, with the aim of correctly identifying the source of the defects to address them and prevent them from recurring. 

The typical approach used, in line with the principles of the lean method, involves’operational process analysis, historical recording and subsequent classification of defects, and the systematic targeting of defect sources. Expected results include a measurable impact on quality-related indicators and the creation of a set of process standards useful for maintaining zero-defect conditions.

The Quality Control process is divided into 7 steps: 

1. Analysis of the manufacturing process and sources of initial defects. 
2. Initial restoration of defect sources and definition of the control network. 
3. Historicization and classification of chronic defects. 
4. Attack of chronic defect sources. 
5. Definition of Zero Defects Conditions – Standard. 
6. Definition of methods for maintaining conditions for zero defects. 
7. Improving methodologies that guarantee conditions for zero defects. 

A crucial element in this process is the standardization. Standardization (Step 5) means establishing and implementing standards for work methods. If a standard exists but is not used at all, it means it's located in the wrong place, doesn't explain how to do the work correctly, cannot be followed as described, or produces a defect when followed correctly. If no standard exists, operators perform the work their own way, resulting in fluctuating conditions and therefore, changing quality. When preparing work standards, it is important to ensure that the work procedures are adequate, the standards are expressed in specific and concrete terms, and the norms are easily understandable, also through extensive use of diagrams and charts. 

Quality Control and the management process

Quality control, in line with the lean methodology, also includes an anomaly management process, in which the detection of defects triggers immediate containment:

• Contain the error to this single process. 

• Isolate the defective batch. 

• Apply containment to similar at-risk processes. 

• Start troubleshooting. 

• Pursue Poka Yoke (error-proofing) processes. 

This implies that the operator detects the problem and calls the foreman, the foreman responds to the problem, and the process restarts (or continues). Once the problem is resolved, the location, time, reason, etc., are recorded, and changes are reflected in the standard work documents. Countermeasures are sought for small problems and a report is drawn up on short-term fixes and corrective actions for larger problems.  

The pillars of Predictive Quality Control (PQC)

The Predictive Quality Control (PQC) It is an emerging category of industrial Artificial Intelligence solutions, which provide manufacturers with the means to significantly reduce process-driven quality losses and waste, by quickly identifying the root cause and preventing such losses before they occur. PQC notifies operators and supervisors of potential quality failures with sufficient lead time to proactively resolve issues. 

The pillars of PQC are: 

• Live AlertThe system can predict quality issues in real-time by identifying and monitoring process conditions that previously led to defects. 

• ForecastReal-time and historical production data from machines, ERP, MES, and quality systems are analyzed to detect trends. 

The advantages of Predictive Quality Control

The benefits of Predictive Quality Control are numerous: 

• Predict quality issuesPrevent quality failures with real-time alerts that allow operators to proactively adjust process parameters to increase first-pass yield and maintain compliance. 

• Optimize material usageoptimize material usage by predicting scrap rates based on real-time production conditions and alerting supervisors if desired rates are about to be exceeded. 

• Quickly isolate defectsquickly isolate defects by identifying when and where in the production process they occurred, to limit the overall number of scrapped parts. 

• Reduce costsreduce costs by minimizing waste, reducing production variability to limit material waste, and improving labor efficiency. 

The essential functions of Predictive Quality Control

Predictive Quality Control, in perfect synergy with the lean method, involves various company functions: 

• Operatorsthe ability to better manage information through predictive quality analysis, allowing them to prioritize prevention actions. 

• Supervisori/CapirepartoPossibility of having greater operational visibility on the quality of the production line through interactive dashboards and predictive alerts on quality parameters, scrap, and raw materials. 

• EngineeringAbility to analyze production, interpret and validate supporting evidence: root cause analysis and identification of process optimizations to increase efficiency. 

• QualityThe possibility of accelerating product testing by analyzing results faster and communicating with stakeholders more efficiently. 

Integration with Lean Thinking 

Predictive Quality Control is intrinsically aligned with the principles of the Lean method. The Lean philosophy focuses on the elimination of waste (Muda) and the creation of value for the Customer, drastically reducing production lead times. PQC contributes to this goal in several ways:

• Waste reductionBy identifying and preventing defects early on, PQC reduces scrap, rework, and the costs associated with the Cost of Poor Quality. This is a fundamental pillar of Lean, which aims to eliminate any activity that does not add value. 

• Continuous improvement (Kaizen)Real-time data analysis and problem prediction enable a continuous learning and improvement cycle. The information obtained from the PQC can be used to refine processes, just as Kaizen foresees. 

• Jidoka (Automation with a human touch)The PQC's ability to notify operators and supervisors of potential failures sufficiently in advance is an example of Jidoka. The system detects an anomaly and “stops” (through an alarm or proactive action) to prevent defect propagation, but it is human intervention, informed by data, that resolves the root cause. 

• Flow (Tense Flow)reducing defects and downtime due to quality issues according to the flow, PQC helps maintain a smoother, uninterrupted production flow, avoiding bottlenecks and machine downtime. 

• StandardizationAs we've seen, standardization is crucial for quality control. PQC, by providing data on conditions that lead to defects, can help define more effective work standards and ensure compliance, contributing to maintaining the conditions for “zero defects.”. 

• Customer focusPredictive Quality Control, by preventing defects, ensures that the final product meets or exceeds customer expectations, strengthening their satisfaction and brand loyalty. 

The future of Quality is predictive according to the lean method

Predictive Quality Control is not just an advanced technology; it is a paradigm shift in quality management that enhances the lean method. Abandoning the reactive approach of “check and correct”, companies can adopt a proactive strategy that prevents problems before they arise, significantly reducing resolution lead times. This not only results in significant economic savings and efficiency improvements but also strengthens corporate reputation and customer satisfaction.

the integration of artificial intelligence with various detection and data analysis technologies makes this approach not only possible, but increasingly accessible and effective. The ability to analyze enormous quantities of data in real time opens up unprecedented scenarios for quality control in every industrial sector, allowing for the optimization of production lead times according to the principles of the lean method.

For companies embracing the principles of Lean Thinking and the lean method, Predictive Quality Control is not an option, but a natural evolution. It is the tool that allows for ever-closer approximation of the “zero defects” ideal, optimizing every aspect of the value chain, reducing overall lead times, and ensuring an efficient, high-quality workflow.

To deepen 

If you are interested in further exploring how quality control methodologies, data analysis, and the adoption of new technologies can transform your production processes and lead your company towards operational excellence, you can visit https://www.bonfiglioliconsulting.com/it/