Spec-Driven Development · AI Unified Process · 2026

Requirements-driven,
AI-native software.

Specs at the center. AI handles the rest.
A spec-driven, iterative methodology inspired by the Rational Unified Process. Requirements stay at the heart of every project; AI generates, regenerates, and refactors everything else around them. Built for enterprise.
Apr 28

Spec-Driven Development with the AI Unified Process

Live online workshop. Learn to create system use cases for AI collaboration and generate full-stack Java applications from formal specifications.

April 28, 2026 · 3:00 – 6:00 PM CEST · Zoom · $149 (Early Bird $99)

Register
The Challenge

Why traditional development falls short

Code-centric development leads to maintenance problems, hinders modernization, and causes business misalignment.

The Problem

Traditional development is code-centric. Requirements get outdated, documentation drifts, and when bugs appear, we dig through code to understand what the system was supposed to do.

AI coding tools make this worse by generating code faster without fixing the underlying process problems.

The Solution

AI Unified Process flips this around. Requirements stay at the center, and everything else gets generated from them using AI as the consistency engine.

Iterative Improvement: Through short iterations, specifications, code, and tests improve together. Documentation enables sustainable development and modernization.

Test-Driven Consistency: Tests ensure the system behaves the same regardless of code generation changes, enabling safe refactoring and evolution.

How It Works

Four agile phases

Each phase runs short iterations where all disciplines work together, not in sequence. Phases overlap throughout the project lifecycle.

Requirements → AI Generation → Business Review → Repeat

Inception

  • Business Requirements Catalog
  • Initial stakeholder alignment
  • Test strategy planning
  • Quick iterations and feedback

Elaboration

  • Business Use Case Diagrams
  • Entity Models
  • System Use Case Diagrams with business validation
  • Test case development

Construction

  • Detailed System Use Case Specifications
  • AI-generated application code
  • Unit testing, integration testing
  • Developer review and iteration

Transition

  • User acceptance testing
  • Continuous delivery and stakeholder feedback integration
  • Production optimization
  • Continuous improvement

The end-to-end view

One diagram showing how the four phases interlock, from initial requirements through to production.

AIUP process overview
Click to view full size
Two Modes

Greenfield vs Brownfield

The same methodology adapts to two realities, a clean slate, or an existing system you can't break.

Greenfield

Start from scratch

A Requirements Engineer creates use cases and an entity model. The AI agent generates code and tests directly from those artifacts. The Software Engineer reviews, every artifact traces back to a requirement.

Greenfield workflow diagram
Use Case Entity Model AI Agent Code + Tests
Brownfield

Reverse-engineer what exists

Start from the running system. The Software Engineer reverse-engineers the entity model, use case model, and specifications from the existing code. Software Engineer and Requirements Engineer then review those artifacts together, establishing the spec baseline that future iterations build on.

Brownfield workflow diagram
Existing Code Reverse-Engineer Entity + Use Case Model SE + RE Review
The Differentiator

Enterprise-ready, not just another demo

Most spec-driven development initiatives stop at the prototype. AIUP is battle-tested in enterprise environments and adapted for AI-native workflows.

Governance & Traceability

Every line of code traces back to a business requirement. Audit-ready by design, for compliance reviews, security audits, and regulatory frameworks.

Legacy Systems Welcome

The Brownfield workflow reverse-engineers specifications from running code, so you can modernize the mission-critical systems you can't simply replace.

Scales Across Teams

Requirements Engineers, Software Engineers, and business stakeholders work in parallel across iterations. Clear roles, shared artifacts, no single bottleneck.

Risk-Managed AI Evolution

Test-protected regeneration ensures AI improvements never silently change behavior. Upgrade models, swap tools, refactor freely, without fear of regression.

Knowledge That Outlives Teams

Living specifications survive team changes, onboarding, and reorganizations. The system documents itself, no tribal knowledge, no key-person risk.

See It In Action

AI Unified Process explained

Watch how AIUP transforms development workflows.

Core Principles

Six fundamental principles

Principles that ensure success in agile, iterative development.

→ 01

Requirements-Driven

Specifications drive everything else, not code.

→ 02

AI-Assisted

AI handles tedious work; humans focus on business logic.

→ 03

Iterative Improvement

Specs, code, and tests evolve together through short cycles.

→ 04

Test-Protected

Comprehensive tests ensure consistent behavior during AI regeneration.

→ 05

Stakeholder-Centric

Continuous validation with business users at every iteration.

→ 06

Traceable

Every line of code traces back to a business requirement.

Beyond Determinism

Why perfect specifications miss the point

It's not about perfect specs, it's about iterative improvement.

The Determinism Fallacy

Critics argue AI code generation only works with exhaustive specifications that force deterministic output. This assumes we need perfect requirements upfront.

Reality: Perfect specifications are impossible and unnecessary. The real value comes from iterative improvement.

Our Iterative Approach

Through short cycles, specifications become clearer, AI generation improves, and tests get stronger. Each iteration builds on the previous one.

Key insight: Tests ensure consistent behavior regardless of how the AI generates code. This enables safe evolution and modernization.

How iterative improvement works

  • Start Small: Begin with basic requirements and generate initial code.
  • Test Everything: Create comprehensive tests that capture expected behavior.
  • Refine Continuously: Improve specs based on stakeholder feedback.
  • Regenerate Safely: Tests protect against regression during AI code updates.
  • Document Reality: Keep specifications aligned with what actually works.
Why AIUP

Benefits

Measurable improvements across every aspect of software delivery.

Better Business Alignment

Stakeholders review every artifact, ensuring the system matches actual needs.

Sustainable Development

Living documentation enables refactoring and modernization without losing knowledge.

Safe AI Evolution

Tests protect system behavior while AI improves code generation quality.

Iterative Quality

Specifications, code, and tests improve together through continuous cycles.

Complete Traceability

From business requirement to code line, every connection is maintained.

Tooling

Not just a methodology, real tools

AIUP ships with ready-to-use plugins for Claude Code that automate the entire workflow, from requirements to implementation to testing.

Requirements & Modeling

The aiup-core plugin provides slash commands that guide Claude through creating requirements catalogs, entity models, use case diagrams, and specifications, all stack-agnostic.

/requirements /entity-model /use-case-diagram /use-case-spec

Implementation & Testing

The aiup-vaadin-jooq plugin adds technology-specific skills for database migrations, UI implementation with Vaadin, and automated testing with Vaadin Browserless and Playwright.

/flyway-migration /implement /browserless-test /playwright-test

IntelliJ Plugin

The AI Unified Process Navigator is a standalone IntelliJ plugin that links @UseCase annotated test methods to their UC-XXX-*.md specifications via gutter icons, and back from the spec to the test.

@UseCase UC-XXX-*.md
Explore all tools
Get In Touch

Ready to transform your development process?

AIUP combines the best of Rational Unified Process with modern AI tooling.