Here are some important definitions to help you navigate through this book:

• Artificial intelligence (AI): The overarching field of creating intelligent machines that can perform tasks typically requiring human intelligence, such as problem-solving, learning, and decision-making
• Machine learning (ML): A subset of AI that enables machines to learn from data and improve at a task without being explicitly programmed for every step
• Deep learning: A subset of ML that uses artificial neural networks with multiple layers to identify complex patterns in data
• Generative AI: A type of deep learning model that generates new, original content such as text, images, or code based on the patterns it learned from its training data
• AI agents: A software system that automatically perceives its environment, makes decisions, and takes actions to achieve specific goals with minimal human intervention
• Agentic AI: A group of AI agents that can...

In the following sections, let’s dive deep into some common patterns of failure.

Misaligned goals

Too often, AI and data teams chase after accuracy scores, losing sight of the business goal. Does a better metric always mean better business? Not necessarily.

Take the banking world. An attrition warning system might define the target as account closure. An AI system looks for signals right before someone closes their account. But what about customers who quietly move most of their money out and keep the account open? Focusing only on account closures misses the bigger goal: building trust and a lasting customer relationship.

Or consider e-commerce. Many recommendation systems are set up to maximize the click-through rate (CTR). The result? More users click, at least for a while. But when people see low-quality or irrelevant suggestions, they get annoyed. Eventually, they visit less or stop using the site entirely. In these cases, more clicks...

In this chapter, we have unpacked how so many AI initiatives have failed. Most AI failures are not technical; they come from a misalignment of goals, vague values, or poor integration into the business workflow. Optimizing for technical metrics alone can create solutions that “work” mathematically but hurt the business. AI is probabilistic and iterative; it should not be managed like a fixed-scope IT project.

Real success depends on getting the foundation right: a clear business objective and value hypothesis; a strong data foundation; secured, scalable, and automated AIOps; and shared ownership between business and AI/data teams.

What’s next?

Now that we understand the failure patterns and the unique nature of AI work, we can shift from “what goes wrong” to “how to do it right.”

In the next chapters, we will translate these lessons into a practical, repeatable framework for building and executing an AI strategy...

• Barocas, S., Hardt, M., & Narayanan, A. (2019). Fairness and Machine Learning: Limitations and Opportunities. MIT Press. https://fairmlbook.org/
• Bergstra, J., & Bengio, Y. (2012). Random Search for Hyper-Parameter Optimization. Journal of Machine Learning Research, 13, 281–305. https://www.jmlr.org/papers/volume13/bergstra12a/bergstra12a.pdf
• Breiter, M., & Keane, M. (2022). The State of AIOps Pipelines and Scalability. [Industry Report].
• Cameron, W. B. (1963). Informal Sociology: A Casual Introduction to Sociological Thinking. Random House.
• Chapman, P., Clinton, J., Kerber, R., Khabaza, T., Reinartz, T., Shearer, C., & Wirth, R. (2000). CRISP-DM 1.0: Step-by-step data mining guide. https://public.dhe.ibm.com/software/analytics/spss/documentation/modeler/14.2/es/CRISP-DM.pdf
• Davenport, T. H., & Harris, J. G. (2007). Competing on Analytics: The New Science of Winning. Harvard Business Review Press.
• European...