The Agentic Development Lifecycle — a lifecycle built to handle uncertainty, prioritize behavioral guardrails, and treat deployment as the start of continuous learning rather than the end of a project.
The transition from SDLC to ADLC is the realization that agentic systems violate the core assumption of traditional software engineering: that behavior is fully specified at build time and strictly deterministic. Agents reason, adapt, and act across external tools. Small shifts in context compound into entirely different outputs.
| Dimension | SDLC | ADLC |
|---|---|---|
| Behavior | Deterministic execution (same input = same output) | Probabilistic execution (behavior evolves based on models, tools, and context) |
| Logic residency | Code and config | Hybrid of code, prompts, foundation models, and vector data |
| Success metrics | Functional pass/fail | Accuracy distribution, hallucination rates, token cost, and alignment |
| Feedback loop | Reactive (users report bugs, devs fix them) | Recursive (telemetry feeds directly back into prompt context or auto-generates new intents) |
Commercially, Stages 0 through 1 through 2 through 3 encompass the Evidence Sprint.
| Stage | Name | Core question |
|---|---|---|
| 0 | Preparation & Hypotheses | The raw capture of intent and constraints. Not everything needs an agent. |
| 1 | Scope Framing & Problem Definition | Establishing the human-agent boundaries and compliance limits. |
| 2 | Agent Definition & Architecture | Designing the non-deterministic structure. |
| 3 | Simulation & Proof of Value | The commercial Evidence Gate. You do not pass this without external signal. |
| 4 | Implementation & Evals | Agent-driven coding and baseline verification. |
| 5 | Testing | Probabilistic testing for reasoning, safety, and edge-cases. |
| 6 | Agent Activation & Deployment | Tiered release with strict autonomy guardrails. |
| 7 | Continuous Learning & Governance | Managing probabilistic drift and feeding the loop. |
The raw capture of intent and constraints. Not everything needs an agent.
2–4 hour structured workshop. 10-minute silent assumption dump (Value, Usability, Feasibility, Viability, Operational). Failure premortem. No untagged claims survive.
Meeting-transcription MCP (Whisper/Fathom); Miro/Slack MCP.
Artifact: Raw intent.md, fully tagged. Facilitator validates tagging.
Establishing the human-agent boundaries and compliance limits.
Defining the agent's autonomy boundaries. Explicitly mapping what requires Human-in-the-Loop (HITL) approval. Drafting the initial evaluation criteria.
Artifact: Refined intent.md with explicit human-agent responsibility mapping and operational KPIs.
Designing the non-deterministic structure.
Compressed design session defining the orchestration pattern, token economics (CAPEX/OPEX), and data architecture. Architecture must satisfy Stage 1 bounds without over-building.
Figma MCP or repo-native prototype design via Claude Code.
Artifact: spec.md. Design/Tech Lead review.
The commercial Evidence Gate. You do not pass this without external signal.
Build an Evidence Slice or Vision Prototype. Attach it to a real client touchpoint. Product + Delivery decide: Promote / Iterate / Pivot / Stop. On a Promote, map scope blocks to the enterprise tracker and sync.
Conversation Intelligence MCP (Gong/Zoom); Enterprise Tracker MCP (Jira/ADO).
Artifact: intent.md (cleared) + per-project ESOA + optional Vision Prototype. PO merge sign-off.
Agent-driven coding and baseline verification.
Git worktree isolation. Plan Mode first — Claude writes plan.md. Engineer accepts the plan, execution begins.
Claude Code CLI; CLAUDE.md as the repo's living context engine.
Artifact: plan.md and verified code diffs in isolated branches. Engineer plan acceptance.
Probabilistic testing for reasoning, safety, and edge-cases.
Non-interactive CI eval suite runs 20–50 task scenarios. Test files are read-only during bug-fix tasks — agents must fix code, not weaken the test.
Headless Playwright/Puppeteer MCP; .github/workflows/agent-evals.yml.
Artifact: Verification logs, evaluation pass results. Automated CI pass threshold.
Tiered release with strict autonomy guardrails.
Multi-pass PR review (Bugs → Security → Alignment). Environment autonomy tiers: Dev = Autonomous, Staging = Clean CI + Auto-review, Prod = Named Human Sign-off.
Artifact: REVIEW.md. Human Release Manager sign-off on Production.
Managing probabilistic drift and feeding the loop.
Statistical process control using Western Electric rules. When an agent hallucinates, breaches a boundary, or triggers a negative feedback loop in production, a background agent diagnoses it and writes a new intent back to Stage 0 automatically.
Infra MCP (Datadog); Product Analytics MCP (PostHog/Pendo).
Artifact: Updated bands.yaml, incident records, fresh intent.md pushed to Stage 0 triage queue.