AI changes software engineering

In 30 seconds

More generated code is only progress if platform, tests, architecture rules and software supply chain keep pace. AI-native development shifts effort from typing to specifying, verifying and securing.

Why it matters

When code is created faster than it can be reviewed, risk shifts into reviews, tests and the supply chain. Speed without quality gates produces technical debt more quickly.

Specification-driven development, automated quality gates and provenance of artifacts maintain quality as the volume of generated code rises.

Editorial assessmentThe bottleneck shifts from “writing code” to “taking responsibility for code”. Architecture Fitness Functions and an AI-SBOM are then not overhead, but what makes pace safe in the first place.

The core architecture questions

• How is generated code checked automatically for quality and security?
• Are architecture rules stored as executable fitness functions?
• Do we know the provenance of our artifacts and models?
• How secure are the deployed coding agents themselves?
• Do platform and golden paths keep pace with the speed?

Typical patterns

• Specification-driven Development: specification before generation.
• Quality Gates: automated tests and checks as a mandatory barrier.
• ADRs & Golden Paths: documented decisions and safe standard paths.
• Provenance & AI-SBOM: traceable provenance of code, models and data.
• Architecture Fitness Functions: architecture rules checked automatically.

Common misconceptions & anti-patterns

• Generated code goes into production without review or tests.
• Architecture rules exist only as a PDF, not as a check.
• No provenance information on models and dependencies.
• Coding agents with far-reaching rights without safeguards.

Practical review questions

1. Does generated code pass through the same gates as handwritten code?
2. Are our central architecture rules automatically checkable?
3. Do we have an AI-SBOM for critical systems?
4. Are the rights of our coding agents limited and auditable?

Link to the end-to-end worked example

How is the agent developed, tested and versioned — and how secure is its supply chain?

• Prompts, tool definitions and model versions are versioned and go through reviews and approvals.
• Generated code passes through the same quality gates as hand-written code.
• The provenance of code, models and data is traceable (provenance / AI-SBOM).
• Architecture rules are stored as executable fitness functions, not just as a document.

To the full worked example: service agent →

Link to the reference architecture

This mission concerns the platform layer and the supply chain. It ensures that the systems designed in the other missions are built verifiably and delivered securely.

Primary sources

Change history

• 25.06.2026 — Editorial Structured according to the mission template.
• 24.06.2026 — Editorial (subject matter) Initial publication with primary sources.