Exploring the differences between zero-shot, one-shot, few-shot prompting and higher-level prompt engineering techniques like the CIC method.
At the Canadian AI Advancement Institute (CAIAI), our mission is to help Canadians learn, adapt, and succeed with AI through accessible AI education and training. As AI systems become increasingly integrated into workplaces, educational institutions, and everyday life, it is essential for stakeholders to understand the essentials of prompt engineering and how to design effective inputs for generative AI tools to get reliable, high-quality results.
In this blog, we will explore the differences between zero-shot, one-shot, few-shot prompting and higher-level prompt engineering techniques like the CIC method.
In this blog, we’ll begin by defining prompt engineering and explaining why it matters for stakeholders using generative AI in professional or personal settings. We will then explore three basic prompting strategies: zero-shot prompting, where the model is given clear instructions without examples; one-shot prompting, which adds a single example to guide output; and few-shot prompting, which uses multiple examples to shape style, structure, and tone. Next, we introduce the CIC method (Context, Information, Constraints), a higher-level framework for building prompts that elicit consistent, repeatable and high-quality AI outputs. We’ll then demonstrate the practical applications of CIC prompting by sharing several CIC-style prompts for both text generation and image generation.
Prompt engineering starts with understanding how different prompting styles affect the behaviour of generative AI. At the most basic level, zero-shot, one-shot, and few-shot prompting give you different degrees of control over how the model interprets a task. These methods are especially useful when you are first learning how to communicate with AI tools or when you need to quickly test an idea before investing time in more advanced techniques.
Zero-shot prompting is the simplest starting point. You provide the AI with a clear instruction, but no examples.
For instance:
“Summarize this article in three bullet points for a non-technical audience.”
Here, the model relies entirely on your wording and its existing training. Zero-shot prompts are useful for everyday tasks such as quick summaries, brainstorming ideas, and asking general knowledge questions. However, because there are no examples, the generative AI tool you are using will most likely miss your preferred tone, depth, or structure, especially for nuanced or specialized tasks.
One-shot prompting adds a single example to clarify what an optimal output looks like.
Using the same task, you might say:
“Here is an example of the kind of summary I want…” followed by one sample, then: “Now, summarize this new article in the same style.”
This small addition helps the model better align with your expectations for format and tone. One-shot prompting is effective when the task is moderately complex or ambiguous, such as drafting an email or creating a basic social media post.
Few-shot prompting goes a step further by providing multiple examples (often three to five) that demonstrate the desired style, structure, and level of detail. For instance, you might include several examples of policy summaries or lesson plans in your prompt, then ask the model to create a new one following the same pattern. Few-shot prompting provides your generative AI tool with additional context to infer patterns in the data you provide, helping it better produce your desired output.
While zero-shot, one-shot, and few-shot prompting can be useful, they can still yield inconsistent results. The CIC method (Context, Information, Constraints) offers a more structured approach to prompt engineering that you can apply across a wide range of tasks and AI tools. Instead of relying on trial and error, CIC provides a repeatable framework for designing prompts that consistently generate relevant, reliable outputs.
When creating a CIC prompt, you will consider three segments: context, information, and constraints.
Context is the foundation of the prompt. It sets the stage by defining the AI tool’s role and abilities. Every context segment should begin with an “imagine” statement, such as: “Imagine you are a ___________ (e.g. marketing strategist for a non-profit, HR manager designing a training plan, etc.). This role framing shapes how the generative AI system interprets the rest of your instructions. To strengthen the context, you then describe the task itself.
Information builds on that frame by supplying the background details the AI needs to work effectively. This might include your target audience, industry, objectives, key messages, timelines, and any additional constraints the model should be aware of. For example, you could provide a brief description of your organization, a product overview, or meeting notes (as long as they are not sensitive in nature). If your AI platform allows it, this is also where you reference uploaded materials, including relevant documents, slide decks, images, code and more. In most cases, the more relevant the information, the better the output.
Lastly, constraints determine the final output format. This is where you specify formatting, including length, tone, and any stylistic requirements. For text generation, constraints could include “500–700 words,” “professional but friendly tone,” “use headings and bullet points,” or “write at a Grade 9 reading level.” For image generation, constraints might cover style (e.g., flat illustration, photorealistic), aspect ratio, colour palette, or resolution. Constraints are critical to the practical application of AI because they ensure the output is usable in real-world contexts.
To see how CIC works in practice, it helps to look at end-to-end examples for both text and image generation. These examples show how you can use the same underlying method across different generative AI tools to support your personal and professional use cases.
Example #1: Sales Research Prompt
“Imagine that you are a sales rep for a organization called ___________ (Company Name). Using the Internet, please conduct a review of their website and gather an understanding of their services, service area and organizational structure. I’d like you to create a bullet point list of 10 potential sales verticals that the company can pursue to expand its business. Please do so in a professional manner that anyone could understand.”
Example #2: Workout Planning Prompt
Imagine you are a certified personal trainer and strength coach specializing in developing safe, effective, and sustainable workout plans for busy adults. Your role is to design a practical, easy-to-follow fitness program that balances strength training, cardio, and mobility work. You are supportive, realistic, and focused on long-term progress rather than quick fixes. I am a healthy adult with no major injuries or medical conditions that limit my ability to exercise.
Assume I am comfortable learning basic exercise technique from short written instructions and, if helpful, brief tips on form and breathing. I’d like you to create a 4-week workout plan with:
Example #3: Cartoon Image Prompt
Imagine you are an animation specialist specializing in creating animated comic-book-style images. I want to create an animated cartoon image of a robot wearing a fast-food restaurant hat and apron set on a black background. Animated style, realism, comic style, realistic features, and crisp edges. (Realism mixed with comic style).
Prompt engineering is quickly becoming a core skill for anyone working with generative AI, and structured methods like CIC help turn scattered experimentation into consistent, professional results. At the Canadian AI Advancement Institute (CAIAI), we’re focused on supporting this kind of practical skill-building through high-quality AI education, including prompt engineering education embedded in our programs.
For individuals and teams seeking prompt engineering courses and training in artificial intelligence, the Canadian AI Advancement Institute offers a range of opportunities, including AI education, an Introduction to Artificial Intelligence course, AI workshops, and AI training tailored to institutional, industry, and government needs.
Whether you are beginning your learning journey or preparing your organization for broader AI adoption, CAIAI offers a range of educational pathways to help you learn, adapt, and succeed with AI.
