qventis blog

Reading a vendor release note like a tester | qventis blog

qventis team — Sun, 11 Oct 2026 09:05:12 GMT

Every enterprise vendor publishes release content before an update lands. Oracle has its What's New pages for each quarterly update. Workday publishes feature release notes ahead of each R1 and R2 preview. Salesforce releases long notes for Spring, Summer and Winter. The volume is large and the writing is aimed at administrators deciding what to switch on and buyers deciding what to get excited about.

Oracle Fusion quarterly update readiness checklist | qventis blog

qventis team — Sun, 11 Oct 2026 09:05:12 GMT

Oracle Fusion Cloud ERP, HCM and SCM follow a predictable rhythm. Readiness content is published ahead of each quarterly update. The update is applied to your test environment, and production follows roughly two weeks later. Oracle EPM runs on a monthly cycle of its own.

Reference architecture: release-aware QE | qventis blog

qventis team — Sun, 11 Oct 2026 09:05:12 GMT

A typical test architecture diagram has a code repository on the left, a CI server in the middle and a dashboard on the right. Commits trigger builds, builds trigger tests, tests report results. That picture works for software you build yourself.

Flaky tests: classify failures by cause | qventis blog

qventis team — Sun, 11 Oct 2026 09:05:12 GMT

Every large suite has a list of tests that fail now and then and pass on a second try. The usual response is a rerun policy: retry failed tests once or twice, and count them as passed if they go green. It keeps the pipeline moving. It also teaches the team to ignore failures, and sooner or later a real defect gets retried into a pass.

Where requirements-to-test traceability breaks | qventis blog

qventis team — Sun, 11 Oct 2026 09:05:12 GMT

Traceability is one of the first things an auditor asks for and one of the last things a project team maintains. The usual artifact is a matrix: requirements down one side, test case IDs along the other, and a mark wherever someone linked the two. A full matrix looks reassuring. It often hides the gaps that matter most.

What agentic test automation should and should not do | qventis blog

qventis team — Sun, 11 Oct 2026 09:05:12 GMT

An agent, in the sense most vendors now use, is an AI system that works toward a goal by planning a sequence of actions and calling tools to carry them out. It can open an application, read a page, write a file, run a job and look at the result, then decide what to do next.

Workday R1 and R2 regression without the weekend | qventis blog

qventis team — Sun, 11 Oct 2026 09:05:12 GMT

Workday ships two feature releases a year, known as R1 and R2, plus weekly service updates in between. Each feature release reaches a preview tenant first, giving customers a window to review new features and test their configuration before production is updated.

Human approval gates for AI-generated heals | qventis blog

qventis team — Sun, 11 Oct 2026 09:05:12 GMT

When an application changes, tests break. A button is renamed, a menu moves, a new required field appears. AI is good at noticing these changes and suggesting how the test should adapt. Many tools call this self-healing and apply the fix automatically.

SAP S/4HANA upgrade testing pitfalls | qventis blog

qventis team — Sun, 11 Oct 2026 09:05:12 GMT

SAP S/4HANA Cloud, public edition, moves on SAP's calendar: two releases a year plus regular feature packs. Private cloud and on-premise customers choose when to upgrade, but the work of an upgrade is similar either way. SAP publishes What's New content, simplification items and deprecations. Your team has to work out what they mean for a system that has usually been extended and integrated for years.

Deterministic execution, AI only at authoring | qventis blog

qventis team — Sun, 11 Oct 2026 09:05:12 GMT

There are two broad ways to put AI into test automation. In the first, a model helps people write tests: it drafts steps from a document, suggests wording, proposes a fix when the app changes. The output is a test that a person reviews and saves. In the second, a model reads each step while the test is running and decides, on the spot, what to click and what to type.