Part II: AI-Native Product Discovery and Design
Chapter 6

AI Product Strategy and Portfolio Thinking

In 2024, a startup spent six months and $800,000 building a custom AI writing assistant. Three months later, a Fortune 500 company launched the same capability as a feature in their existing product, undercutting the startup's differentiation. The build/buy/bake decision is one of the most consequential strategic choices in AI product development, yet most teams make it based on intuition rather than analysis. Done right, it creates defensible moats. Done wrong, it wastes resources on commodity capabilities that should have been bought or baked into existing platforms.
Data Strategy, Not Just Feature Strategy

In AI, a new feature often depends on something far less visible: new examples, new corrections, new evidence about where the system fails. The real ceiling often appears elsewhere—the product was never built to expose its own weaknesses in a form the company can learn from. AI products need a data strategy, not just a feature strategy. The key question: What will this product need to learn next, and how exactly will it learn it? A feature is no longer just a user-facing capability; it is also a bet about future data. The strongest AI companies think of their systems as learning systems: products that create value today while deliberately generating the evidence needed to get better tomorrow.

The Tripartite Loop in Strategic Decisions

Strategic portfolio decisions activate all three disciplines simultaneously: AI PM analyzes competitive dynamics, market positioning, and user value to decide which capabilities deserve investment; Vibe-Coding prototypes different strategic approaches quickly to test which create real differentiation versus false moats; AI Engineering assesses technical feasibility, maintenance burden, and integration costs that determine whether a strategy is sustainable.

Chapter 6 opener illustration
AI product strategy requires thinking in portfolios, not single products.
Vibe-Coding in Strategy Prototyping

Vibe coding enables rapid strategy prototyping. Before committing to a build/buy/bake decision, use quick prototypes to test whether your differentiated approach actually works. Vibe code a bake solution to see if it provides the defensibility you expect, or a buy approach to understand its real limitations. Strategy prototyping through vibe coding reveals which strategic bets are worth taking before you invest in implementation.

Objective: Master the AI product strategy toolkit including build/buy/bake decisions and portfolio allocation.

Chapter Overview

Revised from current chapters on AI PM fundamentals. This chapter covers AI product taxonomy, build/buy/bake decisions, value proposition design, and portfolio allocation for AI products.

Four Questions This Chapter Answers

  1. What are we trying to learn? How to make strategic build/buy/bake decisions and how to position AI products in a competitive portfolio.
  2. What is the fastest prototype that could teach it? A build/buy/bake analysis of your most promising AI initiative, mapping where you have defensible advantage versus commodity capabilities.
  3. What would count as success or failure? Clear taxonomy of your AI product portfolio showing which initiatives create differentiation and which are commodity infrastructure.
  4. What engineering consequence follows from the result? Strategic investment should flow to differentiated AI capabilities while leveraging commodity AI for commoditized components.

Learning Objectives

Sections in This Chapter