Benchmark
One app, two delivery models
The paper compares a ShortStay marketplace built as raw AI-generated code with the same app modelled through Buzzy's semantic application platform.
Whitepaper
The Hidden Cost of AI Application Delivery
AI can generate the first working version quickly. This paper looks at what the business has to own after that: code, dependencies, security review, tests, documentation, handover, upgrades, and operational risk.
Maintenance
The cost after the demo
The analysis looks past feature parity to the code, dependencies, security findings, tests, database lifecycle, and operational responsibilities left behind.
Governance
A smaller custom-code surface
The Buzzy model separates application intent from implementation burden, reducing the amount of app-specific custom code an organisation has to govern.
Inside the paper
What the benchmark covers
Raw AI-generated code
Lines of code, dependencies, security audit findings, tests, infrastructure obligations, and portfolio-scale maintenance exposure.
Buzzy semantic delivery
A semantic application definition running on a maintained core engine, with custom work pushed into explicit extension points.
Technical Debt Index
A practical framework for structural, operational, and organisational debt in AI-created business applications.
Portfolio risk
Why the real issue is not one generated app, but the accidental software portfolio created when every team can generate many apps.