Lean Six Sigma: combining Six Sigma quality with lean speed

Summary of Lean Six Sigma

1. 1

   The Fourth Industrial Revolution: How AI is Transforming Manufacturing

   

   00:00

   00:00

   Imagine a struggling aerospace manufacturer with negative earnings suddenly achieving 20% profitability in just 18 months. What changed? Not massive layoffs or capital investments, but the strategic application of artificial intelligence to identify hidden waste. This transformation represents nothing less than manufacturing's Fourth Industrial Revolution - as significant as Henry Ford's assembly line or Toyota's production system. Even Warren Buffett and Elon Musk recognize AI's transformative power, with Musk calling it "the biggest risk we face as a civilization." The competitive landscape is shifting rapidly, with China racing ahead through its "Made in China 2025" strategy. The question isn't whether to implement AI, but how quickly you can do so before competitors gain an insurmountable advantage.

2. 2

   Finding Hidden Waste and Profit Killers

   Start Listening

   Traditional Lean Six Sigma overlooks crucial inefficiencies. At one aerospace company, AI revealed that low-volume parts - just 20% of revenue - accounted for 73% of setup waste. These "insignificant" parts were major profit killers that conventional analysis missed. The company deployed ten AI-guided metrics targeting specific operational issues. Within 18 months, EBITDA improved from -3.6% to +19.5%. Data mining across their 5,000+ products exposed chaotic pricing - one part priced at $22 cost $26 to make and was successfully repriced to $53. Setup time, central to Toyota's success, allows them to profit at one-fourth of General Motors' volume. While Toyota maintains a setup ratio above 4, the aerospace company's low-volume parts averaged just 1.2 - setup nearly matched machining time. Management's solution of year-long batches created illusory short-term profits but poor cash flow, leading to inventory write-offs of 20% yearly revenue. Traditional Toyota methods require substantial investment - $100,000 per machine plus $10,000 engineering per part - viable only for high-volume production. AI's "Generic Setup Reduction" offers a better solution by eliminating common setup functions across all parts without part-specific engineering. By identifying shared tooling needs, AI can sequence production to cut setup times from 6-8 hours to under 2 hours.

3. 3

   From Chaos to Control: The AI Pull System

   Start Listening

   Unlike chaotic "spaghetti flow" manufacturing where products move randomly, the AI Pull system creates controlled flow by establishing optimal work-in-process (WIP) through dynamic caps based on customer needs, setup times, and capacity. The core concept is simple: exits trigger new starts, maintaining consistent WIP and predictable lead times. While traditional Pull systems worked only in high-volume, low-mix environments, AI extends these benefits to complex low-volume, high-mix operations. When disruptions occur, AI automatically reroutes work to minimize impact. In our aerospace case study, production shifted from 10% completion in week one and 52% in the final week to a balanced 25% weekly - eliminating overtime and quality issues. The system achieved a 75% reduction in setup time for low-volume parts, increasing capacity by 10-20% without added costs. Notably, larger neural networks create better optimization opportunities, reversing traditional diseconomies of scale.

4. Chapter 4

   The Evolution of Manufacturing Intelligence

   Each manufacturing revolution has built upon its predecessor while making the previous approach obsolete. The First Revolution replaced muscle power with steam engines. The Second Revolution, exemplified by Henry Ford's Model T, introduced electricity-powered assembly lines - dropping car prices from $850 to $245 while production exceeded 2 million units by 1920. The Third Revolution came from Toyota after World War II. Their production system reduced profitable production volumes by 75% while maintaining variety through Rapid Setup methods. The Fourth Revolution - where we are now - substitutes data for investment dollars through Artificial Intelligence. While Lean Six Sigma struggles with low-volume, non-repetitive Job Shop Manufacturing, AI optimizes part sequencing to minimize setup costs while meeting delivery dates. Companies that led one technological revolution often fail to lead the next, becoming niche players or failing entirely. The pattern has repeated across industries from computation to semiconductors to steel production. The only safeguards are vigilance, technical competence to embrace the next revolution, and humility - qualities that boards of directors must demand during quarterly reviews. Have you noticed how industry leaders often become footnotes in history books? What makes you think your company is immune to this pattern?

   Chapter 5

   Neural Networks: The Brain Behind Manufacturing Excellence

   Neural Networks enable the Fourth Manufacturing Revolution by learning from examples rather than following explicit rules, similar to how human brains learn through experience. They consist of layered, interconnected "neurons" that process information collectively and learn through back propagation - comparing outputs to examples and adjusting to reduce errors. This allows 75% more efficient part sequencing compared to random ordering. Deep Learning advances this further by automatically extracting data features through multiple hidden layers, eliminating manual effort. While Neural Networks excel at specific tasks like job shop optimization, humans maintain advantages in cross-domain thinking and innovation. The complexity is staggering - sequencing just 4 parts from 50 possibilities creates 230,000 potential combinations. Cloud computing handles this complexity in minutes, training Neural Networks to solve new problems as conditions change. The system's effectiveness increases with larger datasets, turning operational complexity from a challenge into an advantage.

   Chapter 6

   Beyond the Factory Floor: AI in Projects and Development

   AI optimization principles extend to project management and product development, with key differences from manufacturing. These processes lack separate setup times, but high task time variability creates exponentially longer cycle times at peak utilization. Scheduling full 40-hour workweeks leads to project delays. With 70% task variation, pushing teams beyond 85% capacity turns a 3-day cycle into a 2-week cycle at 95% utilization. Brooks' Law confirms that adding people to late projects worsens delays. The solution? Schedule with 15% slack time - seemingly inefficient but ultimately cost-effective. Neural networks identify patterns across projects to reduce costs and cycle times. By analyzing historical data, they group common subtasks for expert teams. This creates steeper learning curves than manufacturing - when tasks are 90% similar, the second implementation costs just 10% of the original. Xerox PARC demonstrated this effectively when engineers found 90% of controller code could be reused across ten copier projects. Using specialized teams across multiple projects, they cut development time by 60% and costs by 40% - proving the value of recognizing project similarities.

   Chapter 7

   The Path Forward: Embracing AI as the Third Factor of Production

   Manufacturing faces diseconomies of scale as revenue outpaces profits. With skilled labor shortages and rapidly obsolete training programs, companies must adopt AI alongside traditional factors of labor and capital. Internet commerce, led by companies like Amazon, creates relentless price pressure. Manufacturers must combine AI with Lean Six Sigma to increase flexibility, enable autonomous adaptation, and reduce costs - despite shorter product lifecycles and market volatility. The semiconductor industry demonstrates AI's impact. Growing from $4 billion in 1978 to $400 billion in 2017, Taiwan Semiconductor Manufacturing Company has leveraged AI through Deep Learning to optimize yields while serving diverse customer needs. As Intel's CEO noted, AI will impact virtually every company - you'll either use it or fall behind competitors who do. The question isn't whether AI will transform manufacturing, but whether you'll lead or follow. The future belongs to those who embrace this intelligent revolution.

   Continue Reading