Five examples of using agentic coding to improve software quality, instead of delivery throughput.

Luca Palmieri

Agentic coding (AI-assisted development) can improve software quality rather than just throughput. When code generation becomes cheap, previously uneconomical practices become viable: building custom tooling, prototyping multiple solutions to discover constraints empirically, implementing low-value-per-line abstractions like safe FFI wrappers, and aggressively paying down technical debt. Real examples include custom linters, safe Rust interfaces over C FFI functions, and migrating C code to Rust for better static analysis coverage. The shift enables quality-focused organizations to raise engineering standards by making verification infrastructure investments economically feasible.

Can agentic coding raise the quality bar?

Example 2: Prototype to discover constraints

Example 4: Low value-per-line abstractions