5.2 Manage Change and Skills Evolution
The arrival of AI does not mean the end of testers, but their evolution. Humans move from the role of "the doer" to "the one auditing the machine".
5.2.1 Essential Skills for Testing with AI
The tester of tomorrow (Certified GenAI Tester) must master:
- Prompt engineering: knowing how to speak to the machine (see Chapter 2).
- Critical evaluation: knowing how to detect a hallucination or bias.
- Technical understanding: understanding limits (context window, temperature).
- Ethics and security: knowing how to sanitize data before sending it.
5.2.2 Building Team Capabilities
How to upskill the team?
- Practical training: workshops, internal Hackathons.
- Knowledge sharing: creating a shared "Prompt Library" within the company to avoid reinventing the wheel.
- Peer learning: pairing between an AI expert and a domain expert.
Red thread: MagicFridge
The QA team sets up "GUS Fridays". Every Friday noon, a team member presents a "Fail" (an AI error) encountered during the week and the corrected prompt that solved it. This creates a culture of collective learning.
5.2.3 Evolution of Test Processes
In an "AI-Enabled" organization, processes change:
- The tester's role: he spends less time writing scripts (task delegated to AI) and more time reviewing and validating the strategy.
- The Test Manager's role: she must define an AI usage policy, manage costs (tokens), and ensure the team doesn't become lazy by blindly trusting the tool.
Red thread: MagicFridge
Before AI: the tester spent 4h writing a SQL dataset and 1h testing.
With GUS: she spends 10 minutes prompting data creation, 20 minutes verifying correctness, and she saved 3h30 for high value-added Exploratory Testing.
🎓 Syllabus point (key takeaways)
- New skills: prompting, critical review, data security.
- Prompt library: key asset to capitalize on team knowledge.
- Role evolution: shift from production (writing) to supervision (review and strategy).