In manufacturing, the processes of review, verification, and validation may appear straightforward. A shoe or a safety helmet, for instance, begins with reviewing the specifications: materials, dimensions, and applicable standards. Next, the product is verified through controlled testing to ensure compliance. Finally, it is validated when used under real conditions, proving that it fulfills its purpose: to protect, support, or provide comfort.
Yet it is precisely in that last step, in real-world validation, where the deepest lessons emerge. A shoe perfectly designed on paper can still prove uncomfortable, or wear out faster than expected. Validation reveals that the initial specifications failed to capture the human experience: movement, temperature, the shape of the foot, or even the cultural habits of use.
Turning ergonomics and comfort into measurable variables is one of design’s greatest challenges. Companies that succeed, balancing technical rigor with human insight, gain a strategic advantage: they can improve because they know how to measure without dehumanizing. You cannot improve what you cannot measure, but measuring without losing what makes us human is an art. The challenge lies in transforming ergonomics, comfort, or perception into verifiable elements without stripping design of its essence. Deviation is inherent to every process; perfection does not exist, and precisely because of that, we can pursue it.
International quality standards such as ISO 9000, ISO 9001, ISO/IEC 25010, and ISO/IEC 29119 define review, verification, and validation as essential components of design and development. But in the era of software and artificial intelligence, these concepts expand and grow more complex.
Verifying that an AI model “works” does not guarantee that it behaves ethically or culturally appropriately in the environment where it will be used. A game or a facial-recognition system can be technically flawless and still culturally inadequate if it was never validated against the diversity of its users.
Validation, in this context, becomes both an ethical and a cultural exercise, one that must ask what real impact the product will have and what values its design represents.
Designing with quality means more than meeting specifications; it means anticipating real use and its effects on people and society. As W. Edwards Deming reminded us, “the customer does not know what can be done.” That is why design must also be an act of education and empathy. Artificial intelligence and software development must acknowledge that culture and society shape what is created. There are no universal solutions: a product or algorithm designed for one culture can have a completely different meaning and impact in another. Applications must therefore be specific to their context, and validated in practice, not merely in simulation.
Intelligent validation means never losing sight of the final impact and the cultural context of everything we design. Validation does not close the design cycle; it restarts it. Each real-world validation is an opportunity to revisit specifications, unlearn assumptions, and redesign. Without unlearning there is no true learning; and without learning, validation becomes a mere conformity check.
To validate intelligently is to measure without losing sensitivity, to adjust without losing purpose, and to recognize that technical success only has meaning when it creates human value. Designing, verifying, and validating are three expressions of a single intention: to build trust. And within that perpetual, imperfect pursuit lies the true essence of quality.
César Díaz Guevara
Quality, Strategy and Innovation Consultant
No hay comentarios:
Publicar un comentario