Chapter 3. The AI Product Development Lifecycle
Software products benefit from short development cycles coupled with frequent feedback. The same is true for building with AI. Rapid prototyping, user experimentation, the data and security constraints you may face, and the volatile nature of LLMs in particular make experimentation and reorientation of your product a feature rather than a bug of the AI product development cycle, illustrated in Figure 3-1.
Figure 3-1. A sketch of the development cycle for AI products; the cycle can be repeated hundreds of times before the AI team arrives at a viable product
To ensure that your project stays on track, this cycle consists of several iterations of defining the use case, building a solution, experimenting with it, and evaluating the results, after which the cycle begins again. To help us get comfortable with this way of working, in this chapter I want to take you on a guided tour of the steps in the AI development lifecycle.
An Exemplary Case Study: Building a News Digest App
To understand the AI product development lifecycle, let’s consider the following hypothetical case. An online news platform is experiencing a dwindling number of subscribers. They know they’re up against a powerful competitor: social media channels are increasingly taking over the role of traditional news media by being the first place people go to find out what’s ...
Get LLM Adoption in the Enterprise now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
