Crafting AI Prompts Framework

Frequently Asked Questions

• Why adopt this framework?

  It's essential to ensure coherence and clarity across your organization when it comes to crafting prompts. This framework guides you on what elements to include in your prompts and how to consistently structure them throughout your organization. The goal is to achieve uniformity in communication and efficiency in the processes.

• What is the difference between the Crafting AI Prompts Framework and the CRAFT Framework?

  They are two different frameworks. The Crafting AI Prompts Framework helps people create effective, responsible prompts for Generative AI. It has two parts: IPE (Interactive Prompt Engineering) for text-based interactions with AI, and IVPE (Interactive Visual Prompt Engineering) for creating images and videos with AI.

  The CRAFT Framework is a separate methodology for software engineering in the GenAI era. It helps teams combine human direction with AI agents to build software with quality, security, and purpose.

• What is Prompt Engineering?

  Prompt Engineering is the practice of crafting and refining prompts so that language models, diffusion models, and other AI technologies can handle tasks more effectively. It helps users understand a model’s capabilities and constraints while improving the relevance and quality of its output.

• Who is the Crafting AI Prompts Framework for?

  The framework is for anyone who wants to use Generative AI more effectively and responsibly for writing, engagement, image generation, or video generation. It is useful for people learning Prompt Engineering as well as developers, engineers, stakeholders, and end users who work with AI tools.

• Which AI tools can I use with the Crafting AI Prompts Framework?

  The framework supports user-facing prompt interactions across a range of AI tools. The IPE approach is relevant to text-based tools such as ChatGPT and Microsoft Copilot, while IVPE applies to visual-generation tools such as Midjourney and DALL·E. Choose the framework section that matches the type of output you want to create.

• How do I get started with the Crafting AI Prompts Framework?

  Start with the Introduction to Prompt Engineering to understand the foundations. Then use the IPE Framework for text-based AI interactions or the IVPE Framework for image and video generation. The cheatsheet is a quick reference while you practise.

• Can I use the Crafting AI Prompts Framework at work or in my organisation?

  Yes. The framework helps teams structure prompts consistently, reuse prompt components, and create a shared prompt library. It also encourages attention to non-disclosure, security, privacy, and responsible AI use when prompts are shared or used in workflows.

• How is the Crafting AI Prompts Framework kept up to date?

  First published in 2023, the framework is updated regularly as Generative AI, prompt techniques, and practical considerations evolve. The latest guidance is maintained in the framework documentation.

• Is this the best Prompt Engineering Framework?

  There is no single best framework for every use case; the right choice depends on your goals, tools, and way of working. First published in 2023 and updated regularly, the Crafting AI Prompts Framework stands out through its component-based structure: prompts can be built from reusable parts, shared in prompt libraries, and adapted into reusable instructions for AI agent skills.

  It also covers both text-based and visual prompting, with guidance for responsible use, including model bias and language and cultural nuances. This makes it a strong choice for individuals and organisations that need prompts to be consistent, reusable, and adaptable across countries, languages, and ways of working. Explore the framework documentation to decide whether it fits your needs.

• Is Prompt Engineering still relevant with the latest AI models and agents?

  Yes. More capable models and AI agents reduce the need for prompt tricks, but they still need clear instructions, relevant context, a defined task, an expected format, and boundaries for responsible use. The Crafting AI Prompts Framework provides this structure through its reusable components and its CRAFT, ING, and AI phases.

  For AI agents, these components can become reusable instructions or skills rather than one-off prompts. The framework therefore remains relevant as a practical way to create, validate, share, and improve the instructions that guide AI systems. Start with the framework introduction.

• What is the difference between Prompt Engineering and Context Engineering?

  Prompt Engineering is the broader practice. It covers the complete process of creating effective AI instructions: defining the goal, task, role, format, boundaries, and the context the AI needs. Context is therefore not separate from Prompt Engineering; you cannot create a meaningful prompt without deciding what information the AI should use.

  The industry often uses the term Context Engineering, especially when discussing AI agents, memory, documents, and tools. In the Crafting AI Prompts Framework, we use Context Management to emphasise that this is ongoing work, not something engineered once and forgotten. Start with Context (C), then continuously select, review, update, and validate the context as the task, information, and AI output change. This helps avoid both missing context and overwhelming the model with irrelevant information.

• Are longer prompts better?

  No. A useful prompt contains the right information, not simply more information. Start with a clear goal, relevant context, task, format, and any important boundaries. Add examples or detail only when they help the AI understand what a successful result looks like.

  The framework’s component structure helps you keep prompts focused and reusable. For complex work, separate or reuse components instead of adding every possible instruction to one large prompt. Review the framework documentation and Prompt Engineering techniques for practical guidance.

• Why does AI sometimes give generic, incorrect, or inconsistent answers?

  AI output often suffers when the goal is unclear, the context is missing or irrelevant, the task is ambiguous, the required format is not stated, or the answer is accepted without validation. Models also have inherent limitations: they can produce convincing but incorrect information, known as hallucination and confabulation, and may reflect bias or incomplete knowledge.

  Use the framework to give clear context and instructions, define what a good outcome looks like, and validate the response before relying on it. If the result is weak, refine the relevant component and try again rather than repeating the same request. The Prompt Engineering guide explains these foundations.

• How do I know whether a prompt works reliably?

  Do not judge a prompt from one good result. Define what success looks like, test it with representative inputs, review failures, and improve the relevant components. For important work, include appropriate human review and safeguards.

  This reflects the framework’s validation and improvement phases: validate whether the output meets the goal, then adapt and improve the prompt, context, or format based on what you learn. Reusable components and prompt libraries make successful patterns easier to share and maintain across a team.

• Is the Non-Disclosure clause necessary in a private instance?

  The non-disclosure clause is essential, serving as a reminder to safeguard company data and comply with GDPR and other security protocols. Even in a private environment, it's crucial to maintain these standards. While a private instance provides a controlled setting, the commitment to non-disclosure should remain a top priority at all times for utmost data security and legal compliance.

  Additionally, even if companies restrict their employees to using private instances, they might still use other tools without permission. Some tools offer both private and public versions, which may not differ significantly from each other. Sometimes, it requires logging in with a work account to access the private instance. This must be verified before sharing confidential data. This is yet another reason to always emphasize the "Non-disclosure" component.

• Does this framework also support image creation?

  Yes, this framework focuses on text-based and media interactions. For image interaction, read the IVPE framework section on the framework page.

• Can I use this content for my presentation, blog, or other materials?

  I am delighted when people write about or present the framework and other content from this website. Sharing this knowledge helps reach a wider audience and supports the mission. However, it is imperative that any use of our materials includes a clear and visible reference to craftingaiprompts.org. This applies to all forms of usage, including but not limited to:

  • Articles and blog posts: Ensure you clearly mention and link back to craftingaiprompts.org.
  • Presentations and educational materials: Include visible credits to the website in your slides or handouts.
  • Adaptations and derivative works: While you are encouraged to write about the content in your own words or adapt it to your presentation style, a visible reference to craftingaiprompts.org is mandatory.
  • Please do not copy the complete content verbatim. Instead, interpret it in your own way and present it in a manner that best serves your audience, always giving appropriate credit to our website.

  Please read more in our Content Usage Policy.

  For any queries regarding usage or to seek permission for special cases, please contact us. If you have seen content from this website used without proper reference or behind a paywall, please let us know!

  Thank you for respecting this work and ensuring that proper attribution is given.

• Why does some content require a login, and is everything on the website free?

  Yes. All educational content on Crafting AI Prompts is free to access, and creating an account is free as well. There is no paid subscription required to view content.

  Some sections require you to log in because the website has previously been targeted by large-scale scraping. A great deal of time and care goes into researching, creating, and bringing this material together. Requiring an account helps us protect that work, better understand how the platform is being used, and reduce misuse while keeping the content freely available to genuine learners.

  By creating an account, you agree to our Account Terms and Conditions and Content Usage Policy. These policies explain how the material may be used and why website content or data may not be copied, republished, or collected in bulk without permission. If you would like to use the content in another way, please contact us.

• Who founded the Crafting AI Prompts Framework?

  Jeroen Egelmeers, Currently: Prompt Engineering Advocate at Sogeti Netherlands. Follow me via LinkedIn or send me an email.

• Why is there a certification, and what does it demonstrate?

  The certification was created because organisations asked for a way to verify that people have genuinely mastered the Crafting AI Prompts Framework and Prompt Engineering. It is not intended as a quick or automatic credential: people who do not properly study the framework documentation and the other preparation resources listed on the test details page are unlikely to pass. The test has 20 questions, takes 15 minutes, and requires a score of at least 85%.

  A certificate is valid for one year and can be shared and verified. Retaking the assessment keeps certificate holders current as Prompt Engineering and the framework evolve. A valid certificate therefore demonstrates verified knowledge of the Prompt Engineering material covered by the assessment.

• Fun Fact: Why the hedgehogs?

  The reason you see hedgehogs everywhere on this website is that the lastname of the founder, Jeroen Egelmeers, starts with "Egel", which is Dutch for "Hedgehog". This is why you'll see it everywhere!