Skip to main content

📖 Glossary

The words of the screen, as AI@Test uses them. The names of the model keep their capital on screen (Test case, Test cycle, Test run, Defect); in a sentence, they take their short form: case, cycle, run, defect. Most of these words are also defined in the application, under Help › Glossary, and on the Agreed vocabulary card of the Admin tab.

In-app Help: the guide of each role, writing an importable case, and the glossary (demo data)

TermMeaning
Test caseWhat a test checks, written in Gherkin: a title, preconditions and Given, When, Then steps
Test cycleAn ordered list of cases, tied to a connection, that are run together. A cycle references its cases; it does not copy them.
Test executionThe dated launch of a cycle, numbered EX-…, which freezes its connection and cases and groups their runs
Test runA case played in a test execution, with its verdict, its log and its screenshots
DefectA deviation SAP reported during a run, tracked until it is closed and retested on its original case
ConnectorThe driving technology of a connection, such as SAP GUI (COM); it sets the MCP server of the execution agent
ConnectionThe precise SAP system the runs reach: host, instance, client and language
Test accountThe SAP account the execution agent opens its session with, shared or personal; its secret is never shown again
App userA person who signs in to AI@Test with their Microsoft Entra ID identity, with a role: Admin, Tester or Auditor
Execution agentThe Windows machine that drives SAP GUI for the runs, one session at a time
DriverThe software that drives SAP GUI on an execution agent, in the version the hub pins. A driver of another version stops the cycle on an INFRA freeze, without charging anything to the cases.
HubThe AI@Test server, which keeps the cases, the cycles and their results, and hands each run to an execution agent
LaneThe right to drive the SAP session: one run at a time, the others wait in the queue
CompilationThe first run of a case, played by the AI agent, which writes its trace. For a case recorded from SAP GUI and imported with its trace, the validation of its recorded trace.
TraceThe sequence of SAP actions a compiled case replays, without AI
BarThe checks a trace passes at the end of its compilation before it enters the catalog: its form, the proof of each Then, then two replays without AI that must pass. Not to be confused with the SAP status bar.
ReplayA run that replays the trace of a compiled case, without AI or tokens: free, with a stable verdict
StaleA case whose trace no longer holds because its Gherkin or its connection changed: its next run recompiles it
Recording"Record in SAP GUI": you work in SAP GUI on your recording station, under your own account, and AI@Test captures your actions and draws the case's steps and trace from them, without AI. You write the Thens.
Test dataThe rows of values of a case, entered in Details › Test data: each row ticked in a cycle gives one run
VariableA value of the case written <x>, declared with its base value in the # Variables block, which each test data row replaces
MappingThe dependencies between the cases of a cycle: a case receives a value that a case played before it created
Played without the value of …The run of a case that receives a value from another case through the Mapping, played without it, on its own or in a test execution started without links: it shows in Results, but the state of the case in its cycle does not count it, since only a chained test execution proves the chain
Consistency checkBefore compilation, a read of SAP that writes nothing and checks that the data and configuration the case assumes exist on the connection. Today it covers SAP GUI connections and order-to-cash cases (@module:SD).
ScheduleThe automatic launch of a cycle, every night, every week or after a transport, in replay only
Frozen (INFRA)The status of a test execution stopped on an infrastructure failure, the INFRA verdict, without a defect: nothing is charged to the cases
Non-probativeA result that proves nothing about the case as it is: its Gherkin has changed since, the run did not play the trace that the compilation validated, or the trace it played has since been found faulty
Then anchoredA Then that the trace proves by a reliable reading of SAP: a message key, the value of a field or a read of the database; for a Then that quotes no value, only a message key or a read of the database. "Passed · x/y Then anchored" says how many of the y Thens are anchored.
Weak proofA Then that the trace proves without anchoring it, for example by a text found on the screen, or by a reading shared with other Thens that does not assert the values the Then quotes: it is not anchored
With partial proofA passed run in which a Then is not anchored (weak proof, unproven, or not attached to this Gherkin). It reads "Passed · x/y Then anchored", grayed out, and counts apart under Blocked, outside the passed runs and the pass rate.
Row not replayed by the barThe mark of a passed compilation whose bar replayed two other data rows, never its own, because the trace writes in SAP: its green rests only on the compilation by the agent, it does not fix a defect and is not compared with the previous run
TagA label in the Gherkin of a case, written like @Sales-OTC, kept in Setup with its color, which filters and groups cases. The module of a case, such as SD or MM, comes from its @module: tag.
ArchiveWhere archived cycles and cases go: they leave the lists, and can be read and restored
Test execution or run?

A test execution, numbered EX-…, is the dated launch of a whole cycle; a run is one case played inside it. "Execution" alone names the Execution screen and the activity of running, never a run or a test execution you open or count.