AI-native SDLC, agentic SDLC and AWS AI-DLC: what each term means
Published
AI-native SDLC and agentic SDLC are two names for the same shift: a software development lifecycle with AI built into every stage, from planning to maintenance, while people stay accountable for the decisions that need judgment. Anthropic's playbook treats the labels as interchangeable. AI-DLC (AI-Driven Development Life Cycle) is narrower. It is a methodology AWS published in 2025, with its own phases (Inception, Construction, Operations), rituals and vocabulary.
Short definitions
AI-native SDLC. Anthropic's playbook calls it "a reimagined process that combines the old control objectives with new enforcement." The linear flow becomes a loop with AI at each point, organized in six stages: Plan, Design, Build, Test, Deploy and Maintain. In the playbook, every stage commits an artifact the next stage can read, such as an intent file or a spec. Atlassian uses the same label and puts the weight on accountability: "the SDLC needs to become legible to agents without becoming less accountable to humans."
Agentic SDLC. Anthropic's playbook notes that you'll also hear this shift "called the agentic SDLC, the AI SDLC, or simply agentic software development," and treats the labels as describing the same thing.
AWS AI-DLC. AWS introduced the AI-Driven Development Life Cycle in July 2025, in a post by Raja SP on the AWS DevOps blog, as "an AI-centric transformative approach to software development." It rests on two ideas. The first is AI-powered execution with human oversight, where AI "creates detailed work plans, actively seeks clarification and guidance, and defers critical decisions to humans." The second is dynamic team collaboration. The work runs in three phases:
- Inception: AI turns business intent into requirements, stories and units of work, and the whole team validates them in "Mob Elaboration."
- Construction: AI proposes architecture, domain models, code and tests, and the team settles technical decisions in "Mob Construction."
- Operations: AI applies the context from earlier phases to infrastructure as code and deployments, with team oversight.
Sprints become "bolts," cycles measured in hours or days, and epics become "Units of Work." In November 2025, AWS open-sourced the workflow as Amazon Q Developer rules and Kiro steering files, and says the workflow "records every human action and approval."
Side by side
| AI-native SDLC | Agentic SDLC | AWS AI-DLC | |
|---|---|---|---|
| What it is | A category: the lifecycle redesigned around AI | Another name for the same category | A named methodology from AWS |
| Who defines it | Several vendors (Anthropic, Atlassian and others) | Same | AWS (Raja SP, who leads Developer Transformation Programs) |
| Stages | Anthropic: Plan, Design, Build, Test, Deploy, Maintain, run as a loop | Same as AI-native | Inception, Construction, Operations |
| Where people decide | At approval gates; Anthropic's playbook reserves human review for regulated and critical code | Same | AI "defers critical decisions to humans"; the team validates in Mob Elaboration and Mob Construction |
| Own vocabulary | intent.md, spec.md, hooks as approval gates (Anthropic) | None of its own | Mob Elaboration, Mob Construction, bolts, Units of Work |
| Tied to a tool | Each vendor maps it to its own (Claude Code, Jira) | No | Open-source rules for Amazon Q Developer and Kiro |
Where they agree, and where they differ
All three keep a person at the decision points, and all three redesign the whole lifecycle, not only code writing. They differ on how much code people still read. In its Deploy stage, Anthropic's playbook describes "layers of agentic review with human review reserved for regulated and critical code." The AWS posts describe teams validating plans and reviewing final artifacts; they don't set a rule for who reads the code before production.
Ithaca's position: on its agentic development platform, people on the client's team read the code before it goes to production, and every agent session is logged with the request, steps, model, cost, result and author. Ithaca runs at level 4 on the autonomy scale Dan Shapiro proposed and adds one rule to it: people read the code before production. Level 5, the dark factory, is where nobody reads the code. More in Levels of AI coding autonomy and Dark factory.
Which term to use
- For the general shift, or when looking for playbooks: AI-native SDLC. Agentic SDLC means the same thing.
- For AWS's method, with its phases and rituals: AI-DLC, and say it is AWS's.
- When evaluating a vendor, skip the label and ask three questions. Who approves each stage? Who reads the code before production? What is recorded for each agent session? Our AI governance guide turns those questions into a checklist.
In practice: Dr. Consulta
Dr. Consulta, a network of medical centers, used Ithaca's platform to modernize its subscription system, which had been in production for eight years. The new system was built alongside the old one and switched over in parts, with Dr. Consulta's team supervising the agents and reading what went to production. One representative feature went from two weeks to up to three days; that is an example, not a measurement of the whole backlog. The Dr. Consulta case study has the details, including what the case does not prove.
Book a demo
FAQ
Is agentic SDLC the same as AI-native SDLC?
Yes, for practical purposes. Anthropic's AI-native SDLC playbook says the shift is also called the agentic SDLC, the AI SDLC or agentic software development, and that the labels describe the same thing.
What is AWS AI-DLC?
The AI-Driven Development Life Cycle, a methodology AWS published in July 2025. AI plans the work, asks for clarification and defers critical decisions to people, across three phases: Inception, Construction and Operations.
Is AI-DLC only for AWS tools?
AWS published it as a methodology and open-sourced workflows for Amazon Q Developer and Kiro. The ideas can be applied with other tools, but the official implementations are AWS's.
Does an AI-native SDLC remove human code review?
Not necessarily. Anthropic's playbook reserves human review for regulated and critical code. On Ithaca's platform, people on the client's team read the code before it goes to production.
Sources
- Raja SP, "AI-Driven Development Life Cycle: Reimagining Software Engineering," AWS DevOps & Developer Productivity Blog, Jul 31, 2025. https://aws.amazon.com/blogs/devops/ai-driven-development-life-cycle/
- AWS, "Open-Sourcing Adaptive Workflows for AI-Driven Development Life Cycle (AI-DLC)," Nov 29, 2025. https://aws.amazon.com/blogs/devops/open-sourcing-adaptive-workflows-for-ai-driven-development-life-cycle-ai-dlc/
- AWS, "Building with AI-DLC using Amazon Q Developer," Nov 29, 2025. https://aws.amazon.com/blogs/devops/building-with-ai-dlc-using-amazon-q-developer/
- Anthropic, "The AI-native SDLC playbook," Aug 21, 2026. https://claude.com/blog/the-ai-native-sdlc-playbook
- Claude Academy, "What changes and where to start." https://academy.claude.com/courses/ai-native-sdlc-playbook/introduction
- Atlassian, "How we're evolving Jira for AI-native software development," Jul 15, 2026. https://www.atlassian.com/blog/company-news/ai-sdlc
- Dan Shapiro, "The Five Levels: from Spicy Autocomplete to the Dark Factory," Jan 23, 2026. https://www.danshapiro.com/blog/2026/01/the-five-levels-from-spicy-autocomplete-to-the-software-factory/
