Test automation in practice and research
Design for Testability — why testability is an architectural decision
Testability is not a side-effect of good architecture. It demands deliberate investment in logging, interface design, tracing infrastructure and architectural decision-making. A survey of the patterns that make those investments pay off.
Read more →Keyword-Driven Testing has been an ISO standard since 2024 — what that means for test-case frameworks
ISO/IEC/IEEE 29119-5:2024 has formalised keyword-driven testing as an international standard. What the standard requires, which frameworks structurally align with it — and where the competition plays in a different discipline.
Read more →AI in test automation — taking stock
75 % of organisations call AI testing strategic, only 16 % run it in production. Where the ROI actually sits today — and which pragmatic tools deliver results with no budget.
Read more →Evaluating Robot Framework tests with AI
Sentence embeddings on RF keywords reach F-score 87 % on redundancy detection. LLM-as-a-Judge rates acceptance-criteria coverage at around USD 1.01 per 1,000 tests. How to check RF suites against Jira stories and wikis automatically — and where the limits are.
Read more →ISO 25010 and 25059 in the context of software quality
ISO 25010:2023 defines 9 main quality characteristics with 40 sub-characteristics. ISO 25059:2023 adds 6 AI-specific sub-characteristics. A structured mapping to test types, automation potential and tooling.
Read more →Natural-language tests versus test code
Natural-language test descriptions produce better embeddings, higher traceability and are easier for LLMs to analyse than code. The evidence from over 30 studies — and where code tests remain structurally superior.
Read more →Missing GUI IDs cost real money
At Google, 84 % of pass-to-fail transitions in CI are not real bugs but flakes. Selector instability is the leading cause. Why test IDs are an architecture decision — and how to enforce them in lint, CI and Definition of Done.
Read more →Does a Playwright-versus-Robot-Framework comparison make sense?
Playwright excels at pure web testing, Robot Framework at heterogeneous technology stacks and regulated environments. Three questions, three tables — and a clear recommendation per scenario.
Read more →Robocorp leaves Robot Framework — Not a quality verdict
Robocorp deprecated Robot Framework because Sema4.ai pivoted to AI agents. What that means for RF users — and what it doesn't.
Read more →Robot Framework and LLM code generation
RF's keyword-driven syntax hits the sweet spot for LLMs: natural-language enough to leverage their language ability, structured enough to keep syntax errors low. ~5 rules instead of ~15, ~40 % fewer tokens — and a real Achilles heel in whitespace.
Read more →Robot Framework is a test framework, not a programming language
Why RF should stay logic-free at the test-case level: ISTQB, IEEE, Pekka Klärck, and the literature all converge on this. With Python keywords as the adaptation layer.
Read more →RPA and test automation — related, but not the same
Both click interfaces, fill forms, read results. Yet RPA and test automation solve different problems. Where the line is — and when RPA-based testing makes sense.
Read more →UiPath Test Cloud — platform ambition meets complex reality
UiPath is a Leader at Gartner and Forrester 2025, yet holds only 4.2 % market share in testing. What the platform delivers, where it hits limits, and who it actually fits.
Read more →Why test automation fails in most organisations
70 % of test automation initiatives fail to meet expectations — rarely because of the tools, almost always because of the architecture. Five causes, one negative feedback loop, and what a well-designed architecture does differently.
Read more →The Robot Framework identity crisis
Since RF 4.0, programming-language constructs have grown inside Robot Framework. Seven common community criticisms, the evolution Pekka Klärck has himself acknowledged as tension — and the thin-layer pattern as the pragmatic way out.
Read more →Multi-technology orchestration with Robot Framework
RF's distinctive strength is technology-agnostic orchestration: GUI, REST, SOAP, database, SSH and CLI in one suite, in one report. Where Playwright and Cypress hit limits — and when the RF layer is worth its 30–40 % overhead.
Read more →AI in Software Testing — 27 Use Cases and Their Actual Maturity
Eight production-ready, eleven in early adoption, eight experimental — and 89% of organisations pilot GenAI in QA, but only 16% scale it. A structured survey of every relevant use case.
Read more →AI-native testing tools in regulated industries — why keyword-driven is not yet replaceable
testRigor, mabl and Virtuoso promise AI-augmented testing. For FDA, ISO 26262, DO-178C, IEC 62304 and DORA that is not enough — self-healing collides with determinism. What keyword-driven frameworks structurally deliver.
Read more →Automating SAP Tests with Python
Three approaches to SAP test automation in Python: PyRFC (archived), GUI Scripting via COM, direct ctypes access. State of 2026, with Robot Framework integration.
Read more →Robot Framework as a Pure Domain-Specific Language
Use Robot Framework only as a DSL orchestration layer; write all keywords in pure Python. The pattern that scales beyond 10,000 tests.
Read more →Why test plans often aren't test plans
Many test plans explain the test pyramid instead of the project. What standards actually require — and what makes plans ineffective.
Read more →