Begin with the question closest to the problem you are trying to understand. The directory contains all 100 canonical CKP questions, grouped by their governed themes.
Reliability and Hallucination
- Why does AI make things up?
- How can I stop AI hallucinating?
- Why is AI confidently wrong?
- Why does AI give different answers to the same question?
- How do I know whether an AI answer is trustworthy?
- Why does AI invent sources or quotations?
- Should I ask AI to admit when it does not know?
- Can a better prompt guarantee accuracy?
- Why does AI repeat an error after I correct it?
- When should I verify an AI answer myself?
Context and Relevance
- How much context should I give AI?
- Why does too much context make answers worse?
- Why does AI ignore an important detail I supplied?
- Should I paste my entire document into AI?
- How do I make AI understand what I really want?
- Why does AI answer the wrong question?
- How do I keep background information from overwhelming the task?
- What should go into a project context file?
- How do I handle conflicting information in my notes?
- Can AI decide which context is relevant by itself?
Memory and Continuity
- Why does AI forget what I told it?
- Why do long conversations become less reliable?
- How do I continue a project in a new chat?
- What should AI remember permanently?
- How do I stop repeating myself to AI?
- Is AI memory enough for serious work?
- How do I preserve decisions made across many sessions?
- What is the difference between memory and preparation?
- How do I recover after a project conversation has become confused?
- How can AI maintain continuity without carrying every old message?
Prompting and Instruction
- Why are my prompts becoming longer and longer?
- Do I need to become a prompt engineer?
- What makes a good professional AI instruction?
- Why does a prompt work once and fail later?
- Should I use a standard prompt template?
- How detailed should my instructions be?
- Why does AI follow some instructions and ignore others?
- Should I tell AI what role to adopt?
- Can examples improve AI output?
- What is more important: the prompt or the context?
Professional Work and Quality
- Why does AI produce generic professional writing?
- How do I make AI work like an experienced colleague?
- Why do I spend so long correcting AI output?
- How can AI preserve my voice?
- How do I stop AI changing the meaning of my work?
- Can AI perform expert work without an expert present?
- How do I get consistent output across repeated tasks?
- Why does AI polish weak thinking instead of improving it?
- How should AI handle specialist terminology?
- What does 'human in the loop' actually require?
Decision-Making and Priorities
- Can AI help me decide what to do next?
- Why does AI give me obvious recommendations?
- How can AI prioritise competing work?
- Can AI act as a chief of staff?
- How do I make AI challenge my assumptions?
- Why does AI agree with me too easily?
- How can AI help with strategic planning?
- How should AI deal with uncertainty in a decision?
- Can AI distinguish urgent work from important work?
- How do I prevent AI from creating more work than it saves?
Knowledge Management and Retrieval
- How should I organise knowledge for AI?
- Do I need a vector database?
- What is the difference between RAG and CKP?
- Why does search return relevant documents but poor answers?
- How do I know which source AI should trust?
- Should all company knowledge be available to AI?
- How do I keep AI knowledge up to date?
- What should happen when two authoritative sources disagree?
- How can AI reuse lessons from previous work?
- Why is a folder full of documents not a knowledge system?
Governance, Risk and Accountability
- How do we govern AI without stopping useful work?
- Who is accountable for an AI-assisted decision?
- How do we prevent confidential information leaking into AI?
- When should AI be prohibited from acting alone?
- How can we audit an AI-generated result?
- What should an AI policy contain?
- How do we stop outdated rules guiding new work?
- Can AI governance be automated?
- How should AI handle personal or sensitive data?
- What evidence should be kept when AI contributes to work?
Teams, Organisations and Scale
- How do we stop everyone using AI differently?
- Should teams share prompts?
- How do we keep AI outputs consistent across departments?
- How should we train staff to use AI?
- Why do AI pilots succeed but operational use fails?
- How can organisational knowledge survive staff turnover?
- Can CKP work across different AI systems?
- How do we prevent duplicate AI work?
- How can managers know whether AI is genuinely helping?
- What changes when AI use scales across an organisation?
Architecture, Workflow and the Future of AI Work
- Why is a chatbot not enough for complex work?
- What is the missing layer between people and AI?
- How should an AI workflow begin?
- When should AI retrieve information and when should it generate?
- How do I design AI work that survives model changes?
- Should AI systems prepare information before I ask?
- How can AI move from answering questions to supporting action?
- What makes an AI system professionally dependable?
- Is better AI mainly a model problem?
- What is CKP's central answer to these 100 questions?