Back to the previous chapter III - Understand the capabilities of AI

1. Before anything else, an AI feature is still a feature

<aside> <img src="https://prod-files-secure.s3.us-west-2.amazonaws.com/d8baae53-7dc0-4bad-a65f-26898d6a633d/82692a36-62c4-4887-8d84-9f22fc07a0d3/Silex_Brand_Symbol.png" alt="https://prod-files-secure.s3.us-west-2.amazonaws.com/d8baae53-7dc0-4bad-a65f-26898d6a633d/82692a36-62c4-4887-8d84-9f22fc07a0d3/Silex_Brand_Symbol.png" width="40px" />

AI is the tool among others to solve problems. As usual in product management, start with the problem. AI opens huge new fields of problems to solve and offers new ways to solve existing problems.

</aside>

Building an AI feature is still fundamentally about building a product feature. The best practices of product development remain just as relevant in the world of AI.

AI is a powerful technology that expands the possibilities for innovative product experiences.

However, it's important to remember a few key principles:

• Continue focusing on solving meaningful problems that generate value for your users or customers.
• Assess the ROI (Return on Investment) of your AI feature using clear success metrics, while also accounting for the unique cost structure associated with AI technology.

2. How to select AI use cases

<aside> <img src="https://prod-files-secure.s3.us-west-2.amazonaws.com/d8baae53-7dc0-4bad-a65f-26898d6a633d/cddb3b62-21be-44ec-95fa-cd90539790e0/Silex_Brand_Symbol.png" alt="https://prod-files-secure.s3.us-west-2.amazonaws.com/d8baae53-7dc0-4bad-a65f-26898d6a633d/cddb3b62-21be-44ec-95fa-cd90539790e0/Silex_Brand_Symbol.png" width="40px" /> Don’t get lost selecting too complicated use cases. Focus on “close-ended” use cases that keep the human-in-the-loop.

</aside>

Selecting AI use cases is a critical and potentially risky step for your project. At this stage, there’s a danger of getting sidetracked by “rabbit hole” features that are unlikely to ever reach production.