San Diego, CA – May 28, 2026 – Author Sergey Nosov announced today the publication of Finding What Requirements Hide: AI-Assisted Detection of Ambiguity, Contradiction, and Gaps, a tool-agnostic field manual for the senior practitioner who must catch the defects that natural-language requirements conceal—using artificial intelligence not to draft requirements faster, but to interrogate the requirements a team already has, with discipline and evidence.
Requirements defects are both common and costly. The Project Management Institute, in its Pulse of the Profession research, found that 37 percent of organizations identified inaccurate requirements gathering as a primary cause of project failure, and decades of cost-of-change data show that a flaw that enters the specification and survives into late development or production costs five to ten times more to correct—higher still in safety-critical and regulated systems, where a late defect can force reverification, revalidation, and recertification.
Most requirements still travel as natural language, and natural language was built to hide things. Ambiguity, contradiction, and the silent gap survive every casual read-through, then surface in development or production, long after the cheapest moment to catch them has passed.
Generative artificial intelligence arrived in the requirements engineer’s toolkit as a generator—a faster way to draft user stories and acceptance criteria. Finding What Requirements Hide teaches the harder, more valuable discipline: AI as an auditor that surfaces what the human eye missed. The practitioner treats every model output as a candidate, not a finding, and confirms it against a real stakeholder, scenario, or source before it counts as a defect. As the book frames it, the goal is to put AI to honest work—and to catch it when it lies.
Twelve chapters across three parts move from the linguistic and logical foundations of what hides in requirements, through three audit chapters—one each for ambiguity, contradiction, and gaps—to the disciplines that make the audit reliable in practice: knowing whether it worked, traceability and remediation, governance, and workflow integration. Each audit chapter closes with a verification ritual: a small, disciplined practice that forces the human to do something the model cannot.
Where most current AI guidance for software teams emphasizes generation, this book is built around detection. Its patterns are grounded in linguistics, logic, and the academic literature on requirements quality, none of which change when a new model ships; it names no vendor product and teaches no prompt cookbook, so the practice survives across model generations. It speaks to the practitioners who carry the most consequential requirements work—in medical devices, aerospace, defense, automotive, financial services, energy, and public infrastructure—and to any team for whom being wrong about a requirement is expensive.
Both paperback and Kindle editions are available on Amazon worldwide. The Kindle edition is enrolled in Kindle Unlimited.
Sergey Nosov is a software engineering lead based in San Diego, California, with three decades of experience across finance, government, biotech, energy, and field services—domains where requirements carry regulatory weight, and a defect found late is the most expensive to fix. He is the author of nine books, including the six-book Software Development Principles series, a comprehensive distillation of software design, architecture, and engineering practice, and Code You Did Not Write, on the discipline of technical direction in the age of AI. His latest, Finding What Requirements Hide, carries that same discipline upstream—from the code to the requirements that precede it. It treats generative AI not as a generator of plausible text, but as a disciplined audit partner whose every finding the practitioner must confirm against the real world.
Finding What Requirements Hide: AI-Assisted Detection of Ambiguity, Contradiction, and Gaps
By Sergey Nosov
Format: 6″ × 9″ paperback and Kindle
ISBN: 979-8-1989-5200-3
314 pp.
Published May 28, 2026
Paperback: $19.99 – Kindle: $9.99
For additional materials and inquiries please visit www.orderfactory.com