The ADLC (AI-Driven Development Lifecycle) is OpenAna’s methodology for autonomous engineering. It builds senior-engineer discipline - planning, testing, security, approvals and auditability - into how Ana works, rather than bolting process onto a generic AI tool.

When does Ana ask for approval?

Low-risk code edits proceed normally; medium-risk changes are explained and validated; high-risk infrastructure, production-impacting and destructive operations require explicit approval.

Do teams have to configure the ADLC?

No. The ADLC is built into every Ana engineer, so teams don’t need to set up rules, stages or approval workflows to get the discipline.

Platform · Methodology

The AI-Driven Development Lifecycle.
Discipline built in.

Ana’s ADLC is the methodology every Ana engineer follows: plan, understand, clarify, implement, test, approve, audit and remember - with risk-aware approvals instead of blanket stage gates.

See Ana in action

The architecture

ADLC · live

What the ADLC is

A checklist helps.
An engineer is better.

Most AI lifecycle frameworks bolt process onto a generic tool. Ana’s ADLC is how Ana works by design - no rules, stages or approval flows for your team to configure.

Process overlay on an AI tool

Like handing a junior developer a checklist, a folder structure and a project template.

It creates order and is easier to supervise - but the discipline lives in the checklist, not the developer.

Ana’s ADLC

Like adding a senior engineer who already understands architecture, security, infrastructure, testing and governance.

The discipline is part of how the engineer works - every time, on every task.

Eight phases

One governed loop,
on every task.

Select a phase - or watch the loop run.

Phase 1 of 8

Ana builds an execution plan from the task, the codebase and a risk assessment before writing a line of code - no random generation.
- Task decomposition and scope analysis
- Risk identification and escalation plan
- Dependency mapping across the codebase
- Strategy chosen by complexity

Policy-aware autonomy

Ana asks when it matters -
not at every checkpoint.

Risk-aware approval instead of blanket stage gates: speed where it’s safe, control where it’s needed.

Low-risk code editsProceeds with normal execution

Medium-risk changesStates assumptions and validates first

Security-sensitive remediationApplies policy checks before acting

High-risk infrastructure changesRequires approval

Production-impacting changesRequires governance and approval

Destructive operationsRequires explicit confirmation

Native behaviours

How a senior engineer
works - by default.

Plans before acting

Builds a structured plan from the task context before touching code.

Reads existing code

Understands the codebase, architecture and conventions before changing anything.

Adapts to brownfield

Works within legacy patterns and constraints without manual setup.

Calibrates depth

Adjusts thoroughness to the task’s complexity and risk.

Flags risky actions

Stops for explicit approval on sensitive or destructive operations.

Handles security inline

Applies security, policy and compliance checks as part of the work.

Handles NFRs inline

Weighs reliability, performance, scalability and maintainability as it builds.

Summarizes its work

Explains what changed, why, and what is left.

Takes course corrections

Accepts feedback mid-task and adjusts without restarting.

Maintains memory

Recalls decisions, incidents and history across sessions.

Integrates with your tools

Works natively with repos, scanners, CI/CD, cloud, tickets and observability.

Works across domains

Software, security, SRE/DevOps and AI/ML - not one lane.

One methodology, four domains

The same discipline
in every engineer.

Software Engineer

Development, modernization, testing, reviews and refactoring on large codebases.

Higher-quality first-pass PRs
Fewer review cycles
Consistent, maintainable patterns

Security Engineer

SAST, SCA, IaC, API and cloud findings analysed and remediated with policy-aware fixes.

Findings turned into fixes
Shorter find-to-fix time
Standard remediation across repos

SRE / DevOps Engineer

Infrastructure automation, CI/CD, incident analysis, RCA and configuration drift.

Faster incident analysis and RCA
Pipelines repaired and optimized
Drift detected and corrected

AI / ML Engineer

Data pipelines, model development, MLOps, explainability and responsible AI.

Fewer pipeline failures
Higher MLOps maturity
Models tested and monitored

Integration depth

Works where your
engineering already lives.

Code repositories

GitHub, GitLab, Bitbucket

CI/CD

GitHub Actions, Azure DevOps, Jenkins, GitLab CI

Cloud platforms

AWS, Azure, GCP

Security scanners

SAST, SCA, CSPM, IaC and API security tools

Ticketing

Jira, ServiceNow, Linear

Observability

Datadog, New Relic, CloudWatch, Grafana, Prometheus

Documentation

Confluence, SharePoint, Notion

Identity & access

Okta, Microsoft Entra ID

Data & ML

Databricks, Snowflake, MLflow

Memory, policy and reach make the method real.

The ADLC runs on the Echelon Network - the architecture that gives Ana episodic memory, governance at the Gate, expert routing and integrations across your stack.

Explore the architecture

Echelon · live

Next · 03 Enterprise

Enterprise & deployment

SOC 2 Type II, ISO 27001, HIPAA and GDPR compliant. Run Ana in your own cloud with data residency and a full audit trail.

The core principle

Code generation alone
isn’t engineering.

Engineering needs structure, context, memory, policy, testing and integration. Ana’s ADLC is built around that from the start.

Book a demo