Glossary
The vocabulary of AI-powered QA, without the needless jargon.
Thirty-two terms we use across this site, defined in a sentence or two. Where a word hides a design decision of ours, we say which one.
AI applied to testing
- AI agent
- A program that drives a large language model to carry out a sequence of
actions against a real system, rather than merely producing text. In
SAPFX, agents plan, write and repair tests by
driving a real SAP system over MCP, in a
plan, generate, heal cycle, always under human review.
- AEO (Answer Engine Optimization)
- Optimizing content so it is understood and cited by AI assistants such
as ChatGPT, Claude, Perplexity or Gemini, and not just by classic search
engines. In practice: structured data, self-contained answers and
verifiable facts. It is why this site ships an
llms.txt
file.
- Fine-tuning
- Retraining an existing model on a specific corpus in order to
specialize it. CabrIA applies this locally, on
the customer's real test executions, so the model speaks the vocabulary of
their business without the data ever leaving their infrastructure.
- LLM (large language model)
- A model trained on large volumes of text, able to understand an
instruction in natural language and respond to it. In QA it is used to
generate test cases, explain a failure, or turn a specification into an
executable scenario.
- MCP (Model Context Protocol)
- An open protocol that exposes tools and data sources to an AI model in
a standardized way. SAPFX relies on the
rf-mcp server to let
an agent drive Robot Framework and a real SAP system, instead of guessing
what is happening on screen.
- NLP (natural language processing)
- The set of techniques that let a machine analyze human language. In QA,
NLP is used among other things to match a specification written in prose to
the test cases that cover it.
- On-premise
- Deployment on the customer's own infrastructure, or on isolated
infrastructure in France, rather than in a public cloud. It is the rule at
AI Cabra: test data, source code and logs never leave the customer's
environment.
- RAG (Retrieval-Augmented Generation)
- A technique that combines document retrieval with generation: the model
answers from documents found at question time, not from its training memory
alone. CabrIA uses it with one requirement, every answer cites its sources,
which is what makes verification possible.
- Computer vision
- Automated image analysis. In software testing it makes it possible to
spot a UI element or a visual regression when no reliable technical
identifier is available.
Software testing and QA
- Accessibility tree
- The representation of an interface as exposed by the browser and
assistive technologies: roles, names, states.
CabraFlow resolves its targets there with
getByRole, instead of depending on CSS or XPath selectors that
the slightest layout change breaks.
- Self-healing
- A test's ability to find an element on its own after its
locator has changed. In SAPFX, healing is always
logged, never silent: an invisible correction would turn a test into an
unreliable witness.
- Test case
- The description of a situation to verify: a starting state, actions, an
expected result. It is the basic unit of QA, and what AI helps produce and
then maintain.
- CI/CD
- Continuous integration and continuous delivery: the chain that builds,
tests and deploys automatically on every code change. Automated tests only
pay off once wired into it, because that is where they run without anyone
asking.
- Test coverage
- The share of code or business flows actually exercised by the tests.
SAPFX measures 94% coverage at version 0.7.0, with an 85% floor enforced in
continuous integration.
- ISTQB
- The International Software Testing Qualifications Board, whose syllabus
and vocabulary act as the de facto standard for test documentation.
CabraFlow offers a documentation export in that format.
- Keyword-driven testing
- An approach where tests are written with reusable business keywords,
the technical machinery staying in a separate layer. Two consequences: a
functional tester can read and change a test without reading code, and a
technical change is fixed in one place.
- Locator
- The expression that identifies a UI element so a test can act on it. In
SAPFX, locators live in a separate resource layer, never inline in the
tests, which is what makes them fixable without touching the
scenarios.
- Playwright
- An open source browser automation library with automatic waiting and
selection by accessibility role. SapFioriLibrary and CabraFlow both build
on it.
- Robot Framework
- An open source, keyword-driven automation
framework where a test reads as a sequence of sentences. It is the
foundation of SAPFX and one of CabraFlow's three export targets.
- End-to-end test
- A test that walks a complete business process across every system
involved, as a user would. On SAP, a single journey can cross SAP GUI,
Fiori and an OData API, which is
why SAPFX covers all three channels.
- Performance testing
- Measuring how an application behaves under load: response times,
behaviour as load rises, bottlenecks. AI Cabra offers performance testing
whose analysis is AI-assisted.
The SAP ecosystem
- ABAP
- SAP's own programming language, in which most business code is written,
standard and custom alike. SAPFX scenarios are validated against a real
ABAP platform (A4H), not a simulation.
- ECC and S/4HANA
- The two generations of SAP ERP: ECC, the legacy one, and S/4HANA, the
current one. SapEccLibrary drives both through
SAP GUI Scripting, which matters during
the long years a company runs both side by side.
- Fiori
- SAP's design language for its modern web interfaces, succeeding the
desktop client screens. A Fiori application is technically a
SAPUI5 application.
- OData
- A standardized REST API protocol, used by SAP to expose its data to
Fiori applications and third-party systems. SapApiLibrary covers OData v2
and v4.
- RFC (Remote Function Call)
- SAP's legacy remote function call mechanism, predating web APIs.
SapApiLibrary supports it optionally, for the cases where OData does not
expose what the test needs.
- SAP GUI Scripting
- The automation interface built into the SAP GUI desktop client, which
allows screens to be driven programmatically. It is the mechanism
SapEccLibrary rests on. It must be enabled both server side and on the
client workstation, which is usually a project's first step.
- SAPUI5
- SAP's JavaScript library that powers Fiori applications. Its element
identifiers change between runs, which defeats generic web tools:
SapFioriLibrary is proven compatible from UI5 1.60 to 2.0.
Our building blocks
- SAPFX
- Our open source SAP test automation ecosystem for Robot Framework,
licensed Apache 2.0 and distributed on PyPI as
robotframework-sapfx. It covers three channels, SAP GUI, Fiori
and OData APIs, with no licence fee and no vendor lock-in.
The overview page.
- SapEccLibrary, SapFioriLibrary, SapApiLibrary
- The three SAPFX libraries: SapEccLibrary for the SAP GUI desktop client
(ECC and S/4HANA), SapFioriLibrary for Fiori and SAPUI5 web on top of
Playwright, SapApiLibrary for OData v2 and v4 APIs, with optional RFC.
- CabrIA
- Our R&D programme: an AI model expert in Quality Engineering,
fine-tuned locally on the customer's real test
executions and paired with a RAG knowledge base where
every answer cites its sources. 100% on-premise, test data stays in.
Status: R&D under construction, no figure announced that has not been
measured. The overview page.
- CabraFlow
- Our R&D programme for a
keyword-driven test orchestrator on top of
Playwright. The user demonstrates their process while narrating it aloud,
the test is written in short business phrases and replays on three targets:
native runner, Playwright export and Robot Framework export. Transcription
and reconciliation are 100% local. Status: R&D in design phase.
The overview page.
A question that goes beyond a definition? The FAQ goes into detail.