The landscape of product development has undergone a profound transformation, moving away from a traditional output-centric "project model" towards an outcome-focused "product model," a shift dramatically accelerated and redefined by the advent of generative artificial intelligence (AI). For decades, organizations primarily relied on a sequential, feature-driven approach where executives and stakeholders dictated a prioritized roadmap of features and projects. In this "project model," a designated feature team product manager would craft detailed specifications, designers would create corresponding designs, and engineers would then build precisely to those specifications. This methodology, while providing a sense of control and predictability, often prioritized the delivery of predefined features over the achievement of measurable customer and business outcomes, leading to what many now term a "feature factory" capable of producing numerous products and features that ultimately fail to resonate with users or deliver significant value.
A Historical Perspective: From Output to Outcome
Historically, the project model found its roots in manufacturing and construction, industries where predictability, adherence to blueprints, and strict timelines were paramount. When applied to software development, this often manifested as the Waterfall methodology, characterized by distinct phases: requirements gathering, design, implementation, testing, and maintenance. While offering structural clarity, this rigid approach proved ill-suited for the dynamic and inherently uncertain nature of software, where user needs evolve, technologies change rapidly, and market feedback is crucial for success. The late 20th and early 21st centuries saw the emergence of Agile methodologies, which sought to address the inflexibility of Waterfall by promoting iterative development, customer collaboration, and responsiveness to change. However, even Agile, when improperly implemented, often devolved into a faster project model, where teams delivered features more rapidly but still without sufficient validation of their actual impact. The focus remained on "output" – shipping code – rather than "outcomes" – solving real problems and generating business value.
The cost of software delivery, encompassing everything from coding to deployment and infrastructure, has plummeted over the past two decades, a trend significantly amplified by cloud computing, DevOps practices, and increasingly, AI-powered development tools. This reduction in delivery cost has fundamentally altered the bottleneck in product development. What was once a challenge of simply building and shipping code efficiently has now become a challenge of knowing what to build. This realization has cemented the imperative for the "product model," an approach centered on continuous discovery and validation.
The Product Model: Embracing Discovery and Empowerment
The product model represents a paradigm shift, emphasizing desired outcomes rather than predefined outputs. At its core, it involves empowered, cross-functional product teams working autonomously to identify critical problems, deeply understand user needs, and then discover and validate solutions that address both customer pain points and strategic business objectives. This process is inherently iterative and evidence-based, where teams prioritize learning over simply building. Instead of receiving a list of features to implement, product leaders or stakeholders articulate the overarching problems to be solved or the key outcomes to be achieved. The product team then takes ownership of discovering the most effective solution.
A hallmark of the product model is its relentless focus on discovery, which involves continuous user research, experimentation, and rapid prototyping. It’s not uncommon for these empowered teams to generate 10-20 or more prototypes, or iterations of prototypes, in a single week. This practice, well-established even before the current generation of generative AI tools, has been made significantly easier and faster with modern prototyping platforms like Figma. The goal of this rapid prototyping is not to create a polished, shippable product, but rather to quickly test hypotheses, gather feedback, and iterate on potential solutions, thereby reducing the risk of building something nobody wants or needs. Industry data consistently supports the efficacy of this approach; studies by organizations like the Product Management Institute suggest that companies employing robust product discovery practices see significantly higher product success rates, improved customer satisfaction, and a more efficient allocation of development resources, often reducing time-to-market for validated solutions by 20-30%.
Generative AI: Accelerating Both Learning and Earning
The advent of generative AI has fundamentally reshaped both product discovery and delivery, acting as a powerful catalyst for the product model. On the delivery side, AI tools like GitHub Copilot, Claude Code, and Cursor have dramatically accelerated the coding process, automating boilerplate code generation, suggesting improvements, and assisting with debugging. This efficiency further reduces the cost and time associated with actually building a solution, making the delivery phase faster and less resource-intensive than ever before. This rapid building capability, however, underscores the critical importance of ensuring that what is being built is indeed the right solution.
Crucially, generative AI has also revolutionized product discovery, particularly in the realm of rapid prototyping. What previously required significant effort from designers and engineers can now be achieved with unprecedented speed by product managers themselves. AI-powered design tools can generate multiple UI layouts, variations, and even functional prototypes from simple text prompts, allowing teams to explore a wider solution space much faster. This enables the creation of "live-data prototypes" – functional prototypes that can be put in front of select users and customers to collect real usage data much earlier and at a fraction of the traditional cost. This ability to gather authentic behavioral insights from functional prototypes is a game-changer for the "build to learn" philosophy, enabling quicker validation of value and usability. Furthermore, the speed of these new tools allows product teams to test multiple approaches in parallel rather than sequentially. Instead of optimizing one potential solution, teams can now quickly develop several distinct prototypes, each exploring a different hypothesis for solving a problem, test them simultaneously, and then converge on the most promising direction, significantly compressing the discovery cycle.
The Evolving Product Manager: From Facilitator to Creator
This profound shift from a project to a product model, amplified by AI, necessitates a dramatic evolution in the role of the product manager. The traditional "feature team product manager," whose primary responsibilities often revolved around project managing, writing detailed specifications (PRDs), and facilitating communication, is increasingly finding their value proposition diminished. Their old role, while important for coordination, does not provide the strategic impact required in an outcome-driven environment where the bottleneck is discovery.
Product coach Jeff Patton famously coined the phrase "build to learn vs. build to earn" to delineate the distinct purposes of building in discovery versus delivery, a distinction that resonates powerfully in the age of generative AI.
Build to Learn: The Engine of Discovery
In product discovery, the objective is to build to learn. Here, the product manager, often in collaboration with designers and engineers, is creating artifacts – prototypes of varying fidelity – with the explicit goal of testing critical hypotheses and mitigating the "four big risks":
- Value Risk: Will customers choose to use or buy this?
- Usability Risk: Can users figure out how to use it?
- Feasibility Risk: Can our engineers build what is needed with the available technology and time?
- Viability Risk: Will this solution work for the various aspects of our business (sales, marketing, legal, finance, etc.)?
"Testing" in this context means actively engaging with users and customers to validate value and usability, consulting with engineers to assess feasibility, and collaborating with stakeholders to ensure business viability. With modern prototyping tools, including AI-powered ones, product managers can independently create and iterate on numerous prototypes weekly, directly engaging in this "build to learn" process without always relying on immediate design or engineering support for basic iterations. This hands-on engagement with solution exploration is a cornerstone of the modern product manager’s contribution.
Build to Earn: Delivering Commercial Quality
Conversely, in product delivery, the purpose is to build to earn. This phase focuses on constructing a commercial-quality product that can be effectively sold, serviced, and supported, and upon which customers can reliably run their businesses. The risks in delivery are fundamentally different and encompass a wide array of non-functional requirements: scalability, performance, fault tolerance, reliability, accuracy, privacy, security, operational robustness, provisioning, and internationalization, among others. "Testing" in the delivery phase refers to rigorous quality assurance, ensuring the product meets this extensive list of demands, performs as advertised, and is ready for the rigors of real-world use. While AI assists in accelerating this building process, the core responsibility remains to deliver a robust, high-quality solution.
Cultivating Product Sense: The Unsung Skill
For product managers navigating this evolving landscape, becoming proficient with prototyping tools and discovery techniques is merely the entry point. The true differentiator, and indeed the harder part, is cultivating and honing "product sense." Product sense is the intuitive ability to evaluate learnings from experiments, synthesize diverse data points, understand underlying user motivations, and guide the product’s direction towards a successful outcome. It encompasses a deep understanding of the market, customer psychology, technological capabilities, and business strategy. It’s the skill that allows a product manager to discern which prototypes are most promising, which feedback is most salient, and which strategic pivot is necessary.
Top companies are increasingly re-evaluating their product management interview processes to assess a candidate’s understanding of this "build to learn" paradigm and their ability to demonstrate strong product sense. This often involves scenario-based questions that require candidates to articulate how they would approach discovery, design experiments, and interpret results, rather than merely describe project management methodologies.
Implications for Organizations and Talent
The shift to the product model, underpinned by generative AI, carries significant implications for organizations and the product management talent pool. Companies that embrace this model tend to be more innovative, responsive to market changes, and ultimately more successful in delivering products that customers love and that drive business growth. This requires not just new tools and processes, but a fundamental cultural shift towards empowerment, continuous learning, and an acceptance of iterative failure as a path to eventual success. Leaders must champion this transformation, moving away from command-and-control structures to foster environments where product teams can truly own problems and discover solutions.
For product professionals, this era presents both challenges and unparalleled opportunities. Those who cling to purely administrative or facilitative roles, preferring not to engage in the "building to learn" aspect of discovery, risk becoming obsolete. The "glue" role, while valuable in some contexts, is increasingly automated or absorbed by more strategically engaged product managers. However, for those who embrace the product manager role as a true builder and creator – one who actively designs experiments, crafts prototypes, analyzes data, and applies deep product sense to guide development – this is truly a "golden era." These skilled individuals, equipped with both strategic acumen and hands-on discovery capabilities, are becoming indispensable architects of innovation, driving the creation of truly valuable products that shape markets and enhance lives. The future of product development belongs to those who can build to learn, learn to build, and consistently translate those learnings into tangible, market-leading outcomes.
