Generative AI and LLM Application Testing

Guided briefing

Generative AI and LLM Application Testing video briefing

A focused explanation of chapter 10, turning the AI testing theory into concrete validation checks.

Briefing focus

Module opening

This is a structured lesson briefing. Real video/audio can be added later as a media source.

Estimated time

9 min

1. 1Module opening
2. 2Learning objectives
3. 3Mind map
4. 4Scenario evidence breakdown

Transcript brief

Teach QA professionals how to test LLM applications, RAG systems, prompt-driven workflows, structured outputs, tools, safety behaviour, and regression quality. The briefing explains why the topic matters, walks through a failure scenario, and identifies the artefacts a tester should produce for evidence and auditability.

Key takeaways

• Connect the AI risk to a measurable test or monitor.
• Document the evidence needed for reproducibility and audit.
• Use the lab or scenario to practise the validation workflow.

Module opening

Teach QA professionals how to test LLM applications, RAG systems, prompt-driven workflows, structured outputs, tools, safety behaviour, and regression quality.

Audience. QA engineers testing chatbots, copilots, RAG products, AI agents, and LLM-assisted workflows.

Why this matters. LLM applications are product systems, not just prompts. Quality depends on retrieval, context, tool permissions, output contracts, safety, monitoring, and human expectations.

ISTQB CT-AI mapping. CT-AI 1.7, 1.8, 8.7, 9.1

Learning objectives

• Explain the core quality risk in generative ai and llm application testing.
• Select practical test evidence that supports an AI release decision.
• Apply the module concepts to a realistic QA scenario.
• Produce a portfolio artifact that can be reused in a professional AI testing context.

Mind map

Scenario evidence breakdown

Visual flow

Topic-by-topic teaching guide

1. LLM App Architecture

LLM products usually combine prompt, model, context, retrieval, memory, tools, UI, and monitoring.

2. Golden Sets and Regression

A golden set is a curated set of prompts, expected qualities, references, and known risk cases.

3. RAG Evaluation

RAG testing checks retrieval relevance, answer groundedness, citation quality, and refusal when evidence is missing.

4. Structured Outputs and Tools

LLMs often produce JSON, commands, or tool calls. Valid syntax is not enough; semantics and permissions matter.

5. Safety and Abuse Testing

LLM systems need tests for prompt injection, data leakage, harmful advice, policy bypass, and unsafe automation.

Practical QA workflow

• Start from the user or business decision affected by the AI system.
• Name the AI asset under test: data, feature pipeline, model, prompt, retrieval index, tool, or full workflow.
• Convert the main risk into observable quality signals and release gates.
• Choose the right oracle: deterministic assertion, metric threshold, metamorphic relation, reviewer rubric, comparison, or production monitor.
• Test important slices, edge cases, misuse cases, and change scenarios.
• Record versions, data sources, thresholds, reviewer notes, and decision rationale.

Test design checklist

• What harm could happen if this AI behaviour is wrong?
• Which users, groups, products, regions, or workflows need separate evidence?
• Which metric or observation would reveal the failure early?
• What is the minimum evidence needed for release, shadow mode, rollback, or rejection?
• Who owns the evidence after the model, prompt, or data changes?

Worked QA example

A tester receives a release request for the module scenario. Instead of asking only whether tests pass, the tester writes three release questions: what changed, who could be harmed, and what evidence proves the change is controlled. The answer becomes a small evidence pack: one risk table, one set of representative examples, one automated or reviewable check, and one release recommendation.

Common mistakes

• Treating AI output as a normal deterministic response when the real risk is behavioural.
• Reporting one impressive metric without slices, uncertainty, or business context.
• Forgetting that data, prompts, model versions, and monitoring are part of the test surface.
• Writing governance language that cannot be checked by a tester.

Guided exercise

Use the scenario above and create a one-page evidence plan. Include the decision being influenced, the main risk, the test oracle, the data or examples required, the release gate, and the owner.

Discussion prompt

Which part of an LLM app is most likely to fail quietly: retrieval, prompt, model, parser, tool, memory, or monitoring?

Hands-on lab mapping

• Lab: CourseMaterials/AI-Testing/labs/06_using_ai_for_testing_playwright.ipynb
• Task: Design and automate checks for an LLM workflow, including golden prompts and structured output validation.
• Why this lab matters: it turns the module theory into visible evidence that a release approver can inspect.

Decision simulation

A chatbot answers beautifully but cannot cite sources for high-risk policy claims. Decide whether to release, restrict scope, or block.

Key terms

• RAG: Retrieval-augmented generation, where external evidence is retrieved and supplied to the model.
• Groundedness: Whether generated claims are supported by supplied evidence.
• Golden set: Curated regression examples with expected qualities or references.
• Tool call: A model-initiated action against an external function, API, or workflow.

Revision prompts

• Explain the module scenario in two minutes to a product owner.
• Name three pieces of evidence you would require before release.
• Identify one automated check and one human-review check.
• Describe how this topic changes after deployment.