# Testing Pyramid
Status: public
Confidence: medium (0.725) (verified)
Last verified: 2026-06-01
Generation: ai_assisted


## TL;DR

The testing pyramid is a practical rule for AI coding agents: generate many small, fast checks near the code, fewer integration checks, and a small number of end-to-end checks for critical user flows.

## Core Explanation

Agents can quickly overproduce brittle browser tests because they are easy to describe in natural language. A healthier workflow asks the agent to first add or update unit and integration tests that isolate the changed behavior, then reserve browser tests for user-visible flows that need full-system confidence.

## Source-Mapped Facts

- Martin Fowler describes the test pyramid as a way to balance automated tests with more low-level unit tests than high-level GUI-driven tests. ([source](https://martinfowler.com/bliki/TestPyramid.html))
- The Practical Test Pyramid describes Mike Cohn's original pyramid as three layers: unit tests, service tests, and user interface tests. ([source](https://martinfowler.com/articles/practical-test-pyramid.html))
- Google Testing Blog recommends a pyramid-shaped first guess of 70% unit tests, 20% integration tests, and 10% end-to-end tests. ([source](https://testing.googleblog.com/2015/04/just-say-no-to-more-end-to-end-tests.html))
- Playwright best practices recommend web-first assertions because Playwright waits until the expected condition is met. ([source](https://playwright.dev/docs/best-practices))

## Operational Notes for AI Coding Agents

Ask the agent to state the intended test layer before writing tests. For implementation changes, prefer a failing unit or integration test first. For browser automation, require stable locators, web-first assertions, and a clear reason why lower-level tests are insufficient.

## Further Reading

- [Test Pyramid](https://martinfowler.com/bliki/TestPyramid.html)
- [The Practical Test Pyramid](https://martinfowler.com/articles/practical-test-pyramid.html)
- [Just Say No to More End-to-End Tests](https://testing.googleblog.com/2015/04/just-say-no-to-more-end-to-end-tests.html)
- [Playwright Best Practices](https://playwright.dev/docs/best-practices)