In the era of traditional software, "can we build it?" was the primary constraint enforcing product discipline, but AI has broadly eliminated that constraint. When anyone—engineers, designers, or PMs—can ship full-featured code in hours, the bottleneck to shipping is no longer resourcing, and anyone can be a PM. 

Architecting Coherence addresses the looming crisis in product development: incoherent, fragmented products and PMs who aren’t sure of their roles. As teams ship faster than they can coordinate, products become "janky," trust erodes, and probabilistic AI systems fail in ways that traditional dashboards don’t catch. This book provides a new structural map for the PM role, which encompasses three distinct modes of operation: the Kondo (focused on product taste, discernment, and curation), the Curie (focused on scientific experimentation at scale), and the Wright (the continuous builder who collapses the gap between idea and product).

Readers will move beyond basic "prompt engineering" and learn how to build the evaluative layer and assessment skills necessary to survive a world of automated output. Through post-mortems of public AI failures and practical frameworks for defensive product design, this book teaches PMs how to maintain integrity in systems that are inherently unpredictable.

What You Will Learn

● Effectively operationalize the Three Modes of AI Product Work: Determine when to act as aHow to navigate the transition from Deterministic to Probabilistic Product Management: Learn how to pivot strategy from rigid logic to managing the behavior of unpredictable, generative systems without sacrificing user trust or brand integrity.

● Effectively operationalize the Three Modes of AI Product Work: Determine when to act as a Kondo (curating excellence), a Curie (running scientific evaluations), or a Wright (building directly with AI tools) to maximize your impact in a "pod-less" development environment.

● Frameworks for defensive product design: Gain a toolkit of specific design constraints and "friction points" that guide users safely through AI interactions, ensuring that the product’s implied guarantees never exceed its demonstrated capabilities.

● Identify and mitigate the "Coherence Tax": Discover how to identify the hidden costs of high-velocity AI development and implement structural frameworks that prevent product "jank," feature sprawl, and the erosion of a unified user experience

Who This Book is For

Mid-to-senior level Product Managers and product leads (e.g. Senior PM, Staff PM, Group PM)

Nina has spent the past decade building consequential AI products at some of the world's most scrutinized technology companies, developing firsthand experience and intuition with user and regulatory trust, and shipping into an accelerating and hotly contested product space. At Facebook and Instagram, Nina worked on child safety tooling; at Google, she worked on Google Assistant and Wearables, bringing AI into some of the most personal and ambient contexts in existence, users’ homes and bodies, before moving to Google DeepMind, where she led regulatory compliance and privacy programs for Google Gemini. Prior to Google, Nina worked at Zendesk, building ML systems in customer service. Nina now works at Weights & Biases, building industry-leading MLOps platforms that AI practitioners use to build and maintain AI products and systems in production. Nina’s career experience, mapping across startups, consumer AI and hardware at massive scale, is the driver behind this book's frameworks and insights into the evolution of the PM role.