AI Development Lifecycle
The AI Development Lifecycle, or AI-DLC, is Rory Richardson’s framing for applying AI across the full software development lifecycle, not just code generation: product direction, specifications, implementation, review, modernization, operations, and feedback loops all become AI-assisted and less linear [src-057].
Key points
- AI-DLC compresses the linear SDLC stages that used to flow through tickets, epics, specs, implementation, test, and release [src-057].
- Richardson uses AWS Project Mantle as a reference point: a large Bedrock rewrite estimated at roughly 30 people for a year was delivered in about 76 days with senior developers and about 200,000 lines of non-bloated code [src-057].
- Product management changes because specs that once took weeks can be rewritten in hours, requiring PMs to stay embedded with fast-moving engineering teams [src-057].
- AI-DLC includes the “mucky” parts of development: changing direction, generating and revising specs, creating commits, operating systems, and retiring technical debt [src-057].
- The lifecycle becomes more nonlinear. Epics and ticketing systems still exist, but AI changes the unit of work and the speed at which teams must resynchronize [src-057].
Related entities
Related concepts
- Agentic Engineering
- Human-Agent Collaboration
- Continuous Tech Debt Retirement
- Agent-Native Infrastructure
- Software 3.0
Source references
- [src-057] Amazon Web Services — “The Future of Agentic AI with Rory Richardson | AWS Humans In The Loop Podcast” (2026-05-01)