In the previous lesson, we learned the importance of being an AI Generalist in today’s industries and how an AI Generalist can successfully develop AI products.
In this lesson, we learn about the product development lifecycle. It starts with initialising the product from idea to release. Here, I have listed the steps of the product development lifecycle. In later lessons, I will correlate these steps and concepts to actual development phases.
A product can be tangible or intangible. For example, providing services to clients can also be a product; a pen that writes smoothly is also a product. As an AI Generalist, one must know how the AI product delivers value to its customers.
Stage 1: In this stage, one identifies the pain area. To identify the pain area, product managers generally conduct user interviews about users’ pain points and do a literature review to understand current pain areas. In this stage, product managers stay highly empathetic to understand user pain.
Stage 2: Once the pain areas are collected, the product manager runs sessions with a team of experts or a sample of the interviewed population to gather ideas. These ideas include top-level solutions as well.
Stage 3: The product manager sits with experts to prioritise or sort the ideas. In this case, I personally like a framework called RICE. It prioritises the ideas based on reach, impact and effort. In addition, I compare it with similar products and quantify how this idea can benefit end users. Once the top ideas are selected, they can be merged to form the product description (what it does).
Stage 4: In this stage, the system architecture and the product processes are designed (we will cover this in detail in the upcoming lessons). Domain experts and AI experts can interact with an AI Generalist to develop the product architecture and inherent processes from scratch. In this stage, the AI Generalist plays a crucial role, drawing on their knowledge of processes, products (what value the product delivers), and AI systems. They ensure the AI product is explainable, observable, and reliable (we will cover these concepts in detail in upcoming lessons). In this stage, we also develop product planning and strategies. Product planning includes the channels/platforms where the product will be available, Go-to-market strategy development and a user feedback analysis mechanism.
Stage 5: Once the product architecture is ready, the prototype is developed, and lab testing begins to validate the product. This is called Technology Readiness Level (TRL). After successful prototyping (TRL 6), the MVP (Minimum Viable Product) can be developed (TRL 7). An MVP is a demo product given to users to gather feedback and further improve and develop a full-fledged product ready for production.
Stage 6: After developing the production-ready version, the product is launched through the pre-decided channels so it can reach users easily. Then the Go-to-Market strategy is executed to market the product and turn it into a brand.
Stage 7: In this stage, a proper mechanism is designed that analyses customer feedback on the product. With the help of AI, many innovative mechanisms can be implemented in this stage.
In this lesson, I outline the basic steps I follow to develop a product. Readers can pause here, do some simple research on the terminology mentioned in the lesson (RICE, TRL, Go-to-Market, MVP, etc.), and return to continue with the next lesson. In the next lesson, we will learn about designing architecture and processes in the AI product.
Leave a Reply