For decades, the landscape of product development has been characterized by two distinct methodologies, each with its own philosophy and operational paradigm. While these approaches have coexisted, fundamental shifts in technology, particularly the advent of advanced artificial intelligence, are now dramatically redefining their prevalence and efficacy, pushing the industry toward a new understanding of value creation.
The Enduring Dichotomy: Project vs. Product Models
Historically, and indeed persisting in many organizations even today, the dominant approach has been the project model. This model is fundamentally concerned with output. It typically originates with stakeholders or executives who define a prioritized roadmap of features and projects. For each item on this roadmap, a designated "feature team product manager" meticulously crafts a specification, often in the form of a Product Requirements Document (PRD). This specification then guides designers in creating visual layouts and user experiences, after which engineers are tasked with building the feature precisely to these predetermined specifications. The success metric in this model is often the timely delivery of the specified features, irrespective of their actual impact or market reception. This approach, while seemingly structured, has long been criticized for its inherent limitations, often leading to what industry experts term the "feature factory" syndrome—a continuous output of features that may not genuinely solve customer problems or contribute to strategic business outcomes.
In contrast, the product model centers resolutely on outcomes. This framework begins with product leaders or key stakeholders identifying a significant problem that needs solving. The responsibility then falls to an empowered, cross-functional product team. This team’s initial and crucial mandate is to discover a solution that is truly worth building—one for which there is robust evidence indicating it will deliver the necessary desired outcome. Only once such a solution is validated through rigorous discovery does the team proceed to build and deliver it. A hallmark of this model, even before the widespread adoption of generative AI, has been the rapid iteration of prototypes, with product teams often creating 10-20 or more prototypes per week using tools like Figma to quickly test hypotheses and gather feedback.
A Shifting Bottleneck: From Delivery to Discovery
While both models have a long history, recent technological advancements have drastically altered their relative efficiencies and exposed critical vulnerabilities in the project model. The most profound change has been the dramatic reduction in the cost and speed of delivery. Advances in cloud computing, DevOps practices, open-source software, and most recently, AI-powered code generation tools (such as Claude Code and Cursor) have made the actual construction of features or projects faster and cheaper than ever before. What once represented a significant bottleneck—the engineering effort required to bring a product to life—is now increasingly streamlined.
This newfound efficiency, however, has inadvertently amplified the weaknesses of the project model. Many in the product world now recognize that the project model, when turbocharged by modern delivery capabilities, functions as an exceptionally efficient "feature factory," capable of producing more irrelevant or poorly designed products at an unprecedented rate. The core issue is no longer how fast something can be built, but what is being built.
Consequently, the real bottleneck has unambiguously shifted from delivery to discovery. The critical challenge now lies in identifying and validating a solution that is genuinely "worth building." This entails a solution that addresses both the customer’s needs and the company’s strategic objectives, generates the necessary business outcome, and crucially, solves the problem sufficiently better than existing alternatives to compel customers to switch or adopt. AI, while accelerating delivery, also plays a transformative role in discovery, albeit in fundamentally different ways. It is precisely in this complex domain of product discovery that the nuanced knowledge and strategic acumen of a product manager become absolutely critical.
The Evolving Role of the Product Manager
This paradigm shift presents a significant challenge for many product managers. As organizations transition from the project model to the product model, many product managers find themselves grappling with the nature of their future contribution. The traditional role of project managing, facilitating, and merely documenting requirements—while once valuable—is increasingly seen as insufficient to provide the necessary value in an outcome-driven environment.
Some product managers, especially those with strong technical foundations, have responded by leveraging the capabilities of advanced engineering tools to lean into the building process, taking on more direct engineering tasks themselves. While this path is valid and beneficial for individuals with the requisite skills, it represents only one facet of the broader evolution.
The most forward-thinking product managers recognize that while they are indeed "product builders and creators," their building efforts serve a distinct purpose compared to those of engineers. This distinction was eloquently captured by product coach Jeff Patton, author of User Story Mapping, who coined the phrase "build to learn vs. build to earn" to articulate the fundamental difference between product discovery and product delivery. This framework has gained particular resonance among product teams navigating the product model in the age of generative AI.
Build to Learn vs. Build to Earn: A Deeper Dive
The essence of the modern product development lifecycle lies in understanding and embracing these two distinct modes of building, each with its own objectives, tools, techniques, and definition of "testing."
Build to Learn (Product Discovery):
In product discovery, the primary objective is building to learn. This phase is dedicated to exploring and validating a combination of technology, functionality, user experience, and business constraints to address the four critical risks inherent in discovery:
- Value Risk: Will customers use or buy this? Does it solve a real problem or create sufficient value?
- Usability Risk: Can users figure out how to use it effectively? Is the experience intuitive?
- Feasibility Risk: Can our engineers actually build this within reasonable constraints?
- Viability Risk: Will this solution work for our business? Is it aligned with our strategy, legal, ethical, and financial constraints?
The methods in this phase are characterized by rapid, low-fidelity, and often disposable prototyping. Modern tools, especially those leveraging generative AI, have made it incredibly easy for product managers, often without direct reliance on designers or engineers, to create 10-20 or more prototypes or prototype iterations per week. The main purpose of these prototypes is to test against the aforementioned risks. "Testing" in this context means:
- Value and Usability: Engaging with users and customers to observe their reactions and gather feedback on the proposed solution.
- Feasibility: Collaborating with engineers to assess the technical challenges and potential implementation roadblocks.
- Viability: Consulting with stakeholders across the organization (e.g., sales, marketing, legal, finance) to ensure alignment with business objectives and constraints.
Build to Earn (Product Delivery):
Once a validated solution emerges from the discovery phase, the focus shifts to building to earn. This phase involves developing a commercial-quality product that can be successfully sold, serviced, and supported, and upon which customers can reliably run their businesses. The risks in this phase are fundamentally different and encompass a broader array of operational and non-functional requirements:
- Scale: Can the product handle a large and growing user base?
- Performance: Is it fast and responsive under various loads?
- Fault Tolerance and Reliability: Can it withstand failures and operate consistently?
- Accuracy: Does it perform its functions precisely as intended?
- Privacy and Security: Does it protect user data and comply with regulations?
- Operations: Is it manageable, monitorable, and deployable?
- Provisioning and Internationalization: Can it be easily set up for new users/regions and adapted for global markets?
"Testing" in the delivery phase means rigorously ensuring that the commercial product addresses this extensive list of demands, performs robustly, and works exactly "as advertised." It involves quality assurance, performance testing, security audits, and operational readiness checks.
Impact of Generative AI on Prototyping and Discovery
The advent of generative AI has introduced two critical practical differences in how "build to learn" is executed today:
-
Accelerated Live-Data Prototypes: While the four major types of prototypes (low-fidelity, high-fidelity, functional, live-data) remain in use, AI-driven tools have dramatically reduced the cost and time associated with creating live-data prototypes. These are functional prototypes that interact with real or simulated data, allowing product teams to place product iterations in front of select users and customers much earlier in the cycle. This enables the collection of actual usage data and behavioral insights at a fraction of the traditional cost and time, fundamentally changing the economics of validating product hypotheses. This capability is a significant game-changer for the "build to learn" process, offering unprecedented speed and fidelity in early validation.
-
Parallel Experimentation: Historically, product discovery often involved a largely sequential iteration process. Teams would typically start with what they believed was the most promising approach, iterate on it until sufficient evidence for productization was gathered, and then declare victory and proceed to delivery. However, the speed and ease of creating prototypes with gen AI-based tools now allow for parallel experimentation. It is increasingly common for teams to quickly develop several distinct prototypes, each exploring a different approach to solving the same problem. These prototypes can then be tested simultaneously, providing a broader evidence base and allowing the team to converge on the most promising solution more rapidly and with greater confidence, subsequently refining it through sequential iterations. This parallel approach significantly de-risks the discovery phase and accelerates the path to market-fit.
The Product Manager’s Golden Era: Skill Evolution and Product Sense
This profound shift mandates a re-evaluation of the core competencies required for strong product management. While becoming proficient with new prototyping tools and discovery techniques is increasingly straightforward, the true differentiator and the "hard part" is cultivating product sense. Product sense is the intuitive ability to evaluate learnings from discovery, synthesize disparate insights, anticipate market needs, and guide the product’s direction strategically. It’s about discerning what signals from prototypes and user feedback truly matter, and how to pivot or persevere based on that intelligence.
The interview processes at leading technology companies are already evolving to assess candidates’ understanding of this new paradigm, specifically their ability to engage in building and testing prototypes as part of a continuous learning loop. Product managers are no longer just orchestrators; they are hands-on experimenters and strategic thinkers.
This evolution inevitably creates a divergence in career trajectories within product management. For those who embrace the builder/creator nature of the role, actively participating in the "build to learn" process, developing their product sense, and mastering modern discovery techniques, a "golden era" is unfolding. These individuals will be highly sought after, driving innovation and delivering significant value in an increasingly competitive market.
Conversely, product managers who prefer to remain in a purely facilitative, managerial, or "glue" role—eschewing the hands-on engagement with discovery and prototyping—will find their positions increasingly at risk. The value proposition of such roles diminishes as tools and processes streamline traditional coordination tasks, and the imperative shifts firmly towards tangible outcome generation through validated solutions.
In conclusion, the dual forces of evolving development methodologies and the transformative power of artificial intelligence are reshaping the very foundation of product creation. The distinction between "build to learn" and "build to earn" is not merely semantic; it represents a fundamental recalibration of priorities, processes, and required skill sets. Organizations that embrace this distinction and empower their product teams to prioritize rigorous discovery and continuous learning will be best positioned to innovate effectively, mitigate risk, and achieve sustainable success in the dynamic global marketplace. The future belongs to product managers who are not just managers of projects, but skilled architects of learning and builders of validated solutions.
