16/01/2025
Whichever industry journal you pick up, whichever news site you visit, whichever thought leader you follow—everyone is talking about AI. And not just talking: they’re experimenting, implementing, and investing. At CQM, we believe it’s time for a more critical perspective. Too often, the conversation around AI is one-dimensional, focusing solely on the newest tools and buzzwords. This can lead businesses to invest in technology that doesn’t solve their real challenges.
That’s why we believe it’s essential to take a step back and examine AI from a mathematical perspective—redirecting the focus to more traditional, proven AI techniques. They may generate less hype, but predictive models, simulations, and mathematical optimization have been delivering competitive advantage for decades. In this article, we explain why traditional AI still outperforms generative AI when it comes to supply chain-related challenges.
AI is more than just ChatGPT
The buzz around generative AI—such as ChatGPT—is impossible to ignore. Whether at work or in daily life, everyone is discussing its capabilities: generating text, writing code, or summarizing large volumes of information. In many cases, this saves time. But the most pressing challenges organizations face—such as optimizing supply chain processes—cannot be solved with generative AI.
This is where traditional AI comes into play. It’s a form of technology that has been used successfully in industry for decades, delivering measurable, reliable outcomes.
Fig. 1: Artificial Intelligence is much more than generative AI.
What is generative AI, and what is traditional AI?
Generative AI (Gen AI) refers to AI that creates new content. It’s trained on vast datasets and uses that knowledge to generate text, images, or other outputs. Think of tools like ChatGPT, capable of composing complex sentences or proposing creative solutions. Yet this technology comes with limitations. One of the most discussed issues is “hallucination”—the generation of incorrect or even fabricated information. This makes Gen AI less suited to contexts where accuracy and reliability are critical—such as business decision-making.
Traditional AI (Trad AI), by contrast, focuses on predictive and prescriptive analytics. It uses advanced statistics, Machine Learning, and mathematical optimization to solve concrete problems. This technology is applied in core operational processes such as demand forecasting, route optimization, and inventory management. Built on hard data, Trad AI delivers robust, repeatable, and trustworthy insights—exactly what businesses need to support strategic and operational decisions.
|
|
Traditional AI |
Generative AI |
|
Primary goal |
Perform pre-defined tasks and support decision-making. |
Creating content, generating ideas, answering questions. |
|
Flexibility |
Limited to the trained domain and purpose. |
Can apply knowledge across different domains and adapt to new situations. |
|
Reliability |
Predictable and reproducible outcomes. |
May produce unpredictable results and draw unexpected connections (hallucinations). |
|
Transparency |
Outputs are explainable and traceable. |
Results are difficult to interpret (black box). |
|
Value added |
Optimizes business processes. |
Automating tasks. |
Table 1: A comparison of traditional AI and generative AI.
Why generative AI doesn’t solve everything
Generative AI absolutely has its place—but it’s no silver bullet. In practice, it works best in secondary processes: aiding communication, supporting ideation, or speeding up creative workflows. But when it comes to mission-critical business processes—where consistency and accuracy are non-negotiable—Gen AI quickly reaches its limits. An error in prediction or a misjudged recommendation can have serious consequences.
Imagine you are running a production plant. You need to decide what to produce, when to switch lines, and how to respond to disruptions. There’s no room here for “hallucinations.” You need models that are grounded in real data and use mathematical optimization. One example is the work we did with Fibrant—a case that perfectly illustrates the strengths of traditional AI.
The power of traditional AI in physical supply chain processes
Fig. 2: The supply chain of physical goods—from product development through production, warehousing, and transport.
Traditional AI plays a crucial role in primary supply chain operations. At CQM, we’ve been using Data Science to enhance business performance for decades. Below are a few practical examples of how traditional AI adds real value:
- Product development
Case: Smart and connected shavers for Philips
Smart integration: By linking an app to a physical shaver, users receive real-time advice for optimal use.
Value: Increased customer satisfaction and brand loyalty.
- Manufacturing
Case: Fujifilm’s production planning tool
Minimizing cutting waste: Mathematical optimization reduces material waste, while the planning system improves the coordination of machines and staff.
Value: Significant financial and operational gains, combined with a more sustainable and efficient production process.
- Warehousing
Case: Smart Logistics at Albert Heijn
Improved productivity: Smart container loading and pick order generation reduce time and travel within the warehouse.
Value: Cost savings, greater efficiency, and deeper insight into productivity drivers.
- Transport
Case: Complex planning at Den Hartogh Liquid Logistics
Intermodal transport optimization: A bespoke APS system uses prescriptive analytics to optimize the movement of chemical tanks across Europe—factoring in transport modes, cleaning, repositioning, and repair.
Value: Enhanced planning, better asset utilization, increased productivity, and more proactive strategic decision-making in a complex logistics landscape.
While traditional AI may not generate headlines, these examples show how it consistently delivers value to companies that want to thrive—not just survive—in their markets.
Real strength: combining domain expertise and technology
What truly makes traditional AI powerful is the combination of human insight with technical expertise, process knowledge, and deep domain understanding. At CQM, this integrated approach consistently proves its value. Technology alone isn’t enough—without understanding the problem context, you won’t unlock its full potential.
Our working method brings these elements together (as also explained in the article “Smart warehousing doesn’t have to be expensive thanks to AI” on Logistiek.nl). By collaborating closely with our clients, we develop solutions that are not only technically sound, but also practical and effective.
The right technology for the right challenge
Generative and traditional AI each have their place. Generative AI is a powerful tool for creative and supportive tasks, while traditional AI is the key to solving concrete business challenges. For organizations seeking real impact, the priority is to use the right technology for the right task. In the end, it’s not about what’s trendy—it’s about what delivers tangible results.
Curious how traditional AI could drive impact in your organization?
For over 45 years, CQM has been applying traditional AI to help organizations grow. From idea to implementation, from change to long-term support—we create lasting value for processes, culture, and systems. Interested in what traditional AI can do for your goals? Get in touch with Peter Hulsen for a no-obligation introduction.
