Exploratory-testing charters & personas
The problem
Exploratory sessions drift or repeat the same happy paths.
Workflow
- Feature + risksstory, past bugs
- LLMcharters per persona
- Sessiontime-boxed testing
- Notesfindings
- Summarybugs + new test ideas
Real-world example
For a food-delivery app: charters for 'elderly first-time user on slow 3G', 'user switching between Hindi and English', 'user cancelling exactly as the rider is assigned'. The third finds a refund bug.
Starter prompt
Create 6 exploratory-testing charters for this feature. Each: persona, mission, areas to explore, risks to watch, 45-minute timebox. Prioritise risks from the past defects provided.
Tools
Any LLM, session notes template
Human check
Charters guide humans — the tester's judgement does the finding.
Measure success
Bugs per session; % bugs from exploratory vs scripted testing.