Brief in progress — review the script and demo before recording.
Shaping intent is now mostly the agents’ work. Humans steer the process toward a specific vision here and there. There are far fewer areas where agents cannot handle the work better than we can. We need to show this through a single example, perhaps something you are currently working on in Kody, or something you shipped recently that is valuable for this lesson.
- Intent: explain what intent is, what shape it can take depending on the product, and how we can create it together with AI, considering not only the narrow scope of a given feature but also the wide context of the entire product.
- Context: the narrow and wide context of features and of the product as a whole — how to shape it, iterate on it, and keep it up to date where needed, while preserving decision history and making it manageable across the organization.
- Specifications: agents need not only the narrow and wide context of a project, but often also access to a history of specifications, so they can understand the decisions behind them and how those decisions changed over time. How we organize specifications depends heavily on the project and our needs.
- Evaluation: what can be measured during implementation and after a feature ships? How do we access that data, who should handle it, and who makes the decisions?
We may want to showcase scenarios such as:
- The entire video can showcase a single feature, or a set of features, delivered as a Product Engineer. It would be especially useful to show the entire journey with as many real-life details as we can.