Today, individual enterprise developer productivity depends not only on prompting skills, but to a much greater extent on the effective use of AI-native development assets such as instructions, custom agents, agent skills, saved prompts, MCP server configurations, and more. However, productivity at the organizational level requires further operationalization of AI-native development, including collaborative work on these assets, as well as their distribution and enforcement from the enterprise level.
In this session, I will first demonstrate how these composable AI-native development assets work together, how to map existing task and project knowledge across them, and why each component plays a critical role. I will then show how to accumulate, manage, and provision these assets to developers at the enterprise level to enable reuse, consistency, and standards enforcement. This approach helps break knowledge silos at the code repository level and lays the foundation for an enterprise-wide AI-native development platform.
The session makes a clear case that AI-native development assets become as valuable as code, documentation, and other core project resources, and that today’s tooling already allows them to be treated as first-class artifacts throughout the enterprise development lifecycle.