https://utaa.org/en/blog/ Test automation in practice and research en Fri, 25 Sep 2026 00:00:00 GMT https://utaa.org/en/dossier/robot-framework-as-test-specification-format/ https://utaa.org/en/dossier/robot-framework-as-test-specification-format/ Fri, 25 Sep 2026 00:00:00 GMT 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. https://utaa.org/en/blog/ai-in-testing-strategy-versus-practice/ https://utaa.org/en/blog/ai-in-testing-strategy-versus-practice/ Wed, 02 Sep 2026 00:00:00 GMT 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. https://utaa.org/en/dossier/ai-in-software-testing-use-cases/ https://utaa.org/en/dossier/ai-in-software-testing-use-cases/ Mon, 31 Aug 2026 00:00:00 GMT 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. https://utaa.org/en/dossier/uipath-test-cloud-platform-reality/ https://utaa.org/en/dossier/uipath-test-cloud-platform-reality/ Wed, 26 Aug 2026 00:00:00 GMT 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. https://utaa.org/en/dossier/robocorp-leaves-robot-framework/ https://utaa.org/en/dossier/robocorp-leaves-robot-framework/ Fri, 31 Jul 2026 00:00:00 GMT Robocorp deprecated Robot Framework because Sema4.ai pivoted to AI agents. What that means for RF users — and what it doesn't. https://utaa.org/en/dossier/nl-tests-vs-code-llm-evidence/ https://utaa.org/en/dossier/nl-tests-vs-code-llm-evidence/ Thu, 30 Jul 2026 00:00:00 GMT 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. https://utaa.org/en/blog/keyword-driven-testing-iso-standard/ https://utaa.org/en/blog/keyword-driven-testing-iso-standard/ Tue, 28 Jul 2026 00:00:00 GMT 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. https://utaa.org/en/blog/rpa-and-test-automation/ https://utaa.org/en/blog/rpa-and-test-automation/ Wed, 15 Jul 2026 00:00:00 GMT 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. https://utaa.org/en/dossier/robot-framework-multi-technology-advantage/ https://utaa.org/en/dossier/robot-framework-multi-technology-advantage/ Wed, 08 Jul 2026 00:00:00 GMT 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. https://utaa.org/en/blog/plan-test-ids-from-day-one/ https://utaa.org/en/blog/plan-test-ids-from-day-one/ Tue, 07 Jul 2026 00:00:00 GMT 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. https://utaa.org/en/blog/why-test-automation-fails/ https://utaa.org/en/blog/why-test-automation-fails/ Mon, 29 Jun 2026 00:00:00 GMT 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. https://utaa.org/en/dossier/evaluating-rf-tests-with-ai/ https://utaa.org/en/dossier/evaluating-rf-tests-with-ai/ Fri, 26 Jun 2026 00:00:00 GMT 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. https://utaa.org/en/blog/playwright-versus-robot-framework/ https://utaa.org/en/blog/playwright-versus-robot-framework/ Thu, 25 Jun 2026 00:00:00 GMT 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. https://utaa.org/en/dossier/robot-framework-as-pure-dsl/ https://utaa.org/en/dossier/robot-framework-as-pure-dsl/ Wed, 24 Jun 2026 00:00:00 GMT Use Robot Framework only as a DSL orchestration layer; write all keywords in pure Python. The pattern that scales beyond 10,000 tests. https://utaa.org/en/blog/automating-sap-tests-with-python/ https://utaa.org/en/blog/automating-sap-tests-with-python/ Thu, 04 Jun 2026 00:00:00 GMT Three approaches to SAP test automation in Python: PyRFC (archived), GUI Scripting via COM, direct ctypes access. State of 2026, with Robot Framework integration. https://utaa.org/en/blog/what-a-test-plan-is/ https://utaa.org/en/blog/what-a-test-plan-is/ Fri, 22 May 2026 00:00:00 GMT Many test plans explain the test pyramid instead of the project. What standards actually require — and what makes plans ineffective. https://utaa.org/en/dossier/robot-framework-identity-crisis/ https://utaa.org/en/dossier/robot-framework-identity-crisis/ Mon, 18 May 2026 00:00:00 GMT 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. https://utaa.org/en/dossier/iso-25010-25059-qa-dossier/ https://utaa.org/en/dossier/iso-25010-25059-qa-dossier/ Fri, 15 May 2026 00:00:00 GMT 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. https://utaa.org/en/dossier/design-for-testability/ https://utaa.org/en/dossier/design-for-testability/ Wed, 06 May 2026 00:00:00 GMT 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. https://utaa.org/en/dossier/ai-testing-regulated-industries/ https://utaa.org/en/dossier/ai-testing-regulated-industries/ Tue, 05 May 2026 00:00:00 GMT 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. https://utaa.org/en/dossier/robot-framework-and-llm-code-generation/ https://utaa.org/en/dossier/robot-framework-and-llm-code-generation/ Wed, 15 Apr 2026 00:00:00 GMT 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.