Craig S. Mullins warns of rising data debt in new book

Veteran database expert Craig S. Mullins has released a new book arguing that weak data quality, technical debt and poor governance are now major risks to AI efforts and enterprise decision-making. The book targets leaders building cloud, analytics and AI systems that depend on trustworthy data at scale.

Why it matters: - Mullins argues that AI is amplifying the cost of bad data across enterprises. - Poor data quality can now affect recommendations, automation and decisions, not just dashboards. - The book frames data trust as a core business issue, not just a technical one.

What happened: - Craig S. Mullins announced a new book titled The Cost of “Good Enough” Data: Why Modern Architectures Fail at Scale. - The book examines why modern data architectures often fail despite spending on cloud platforms, data lakes, lakehouses, analytics systems and AI tools. - The release was dated July 1, 2026, from Houston.

The details: - Mullins says enterprises are facing a growing crisis of data debt. - The book says that debt can raise costs, create compliance risks and weaken business decisions. - Common problems in the book include duplicate records, undocumented transformations, fragmented pipelines and weak data lineage. - The book focuses on distributed and hybrid environments, where traditional data management approaches often fall short. - Mullins draws on more than four decades of work in database systems, data architecture and enterprise information management. - The book is aimed at CIOs, CTOs, Chief Data Officers, enterprise architects, database professionals, data engineers, governance leaders and executives. - Mullins says, “Organizations do not have an AI problem. They have a data problem that AI is exposing.” - The book covers the hidden costs of poor data quality and technical debt, AI-driven exposure of data weaknesses, growing pipeline complexity, governance in hybrid environments, trusted data and trustworthy AI, metadata, systems of record and changing roles for DBAs and data architects.

Between the lines: - The book argues that many AI projects fail for reasons that start long before model development. - That shifts attention from buying more tools to fixing the underlying information layer. - The message suggests enterprise data governance is becoming a prerequisite for AI reliability.

What's next: - Mullins is making the book available as a guide for leaders trying to build trusted data foundations for AI. - The release invites contact for additional information, speaking engagements, interviews and review copies. - More information is available on the company's website.

The bottom line: - Mullins is betting that the biggest obstacle to AI success is not the model stack, but the quality of the data underneath it.