The key to consistently generating high-quality outputs from ChatGPT and Google Bard lies in understanding the six essential components of a good prompt: task, context, exemplars, persona, format, and tone. These components work together to provide the necessary information and structure for the language models to deliver tailored and relevant responses.

1. Task: Start the prompt with an action verb (e.g., generate, write, analyze) and clearly articulate the desired outcome. This could be a simple task like creating a training program or a more complex multi-step request.

2. Context: Provide relevant background information to help the model understand the user's situation, desired success criteria, and the environment they are operating in. This helps constrain the endless possibilities and ensures the output is tailored to the specific needs.

3. Exemplars: Including relevant examples or frameworks within the prompt can significantly improve the quality of the output. These exemplars act as a guide for the model to follow, ensuring the response aligns with best practices or a specific structure.

4. Persona: Defining the persona you want the model to embody, such as a subject matter expert or a specific individual, can shape the tone, language, and perspective of the output.

5. Format: Visualize the desired format of the output, whether it's a table, email, bullet points, or a markdown-formatted summary. Specifying the format helps the model structure the response accordingly.

6. Tone: Indicate the desired tone, such as casual, formal, enthusiastic, or pessimistic, to ensure the model's language and voice align with your preferences.

By incorporating these six components into your prompts, you can consistently generate high-quality outputs that are tailored to your specific needs, saving time and effort in the process.

The first and most important component of a good prompt is the task. The rule of thumb is to always start the task sentence with an action verb such as "generate", "give", "write", or "analyze". Clearly articulate what your end goal is, whether it's a simple task like generating a three-month training program or a more complex multi-step request like analyzing user feedback, sharing the top three takeaways, and categorizing the feedback based on the team responsible for follow-up.

The task is the foundation of your prompt, and it's mandatory to include it. Without a clearly defined task, the output from the language model will be generic and less useful. By starting with a strong, action-oriented task statement, you set the stage for the rest of your prompt to be effective.

The second component of a good prompt is context. Context is the trickiest to get right because there is an infinite amount of information that can be provided. However, the key is to provide just enough relevant information to constrain the endless possibilities and guide the AI model to generate a high-quality output.

When crafting the context, it's helpful to ask yourself three questions:

1. What is the user's background? - This helps the AI understand the user's expertise, experience, and any relevant details about them.

2. What does success look like? - Defining the desired outcome or goal helps the AI align its response to meet the user's needs.

3. What environment are they in? - Providing details about the user's current situation or setting can help the AI generate a more tailored and appropriate response.

For example, in the workout program prompt, the context provided is: "I'm a 70kg male looking to put on five kilograms of muscle mass over the next three months. I only have time to go to the gym twice a week and for one hour each session." This context gives the AI the necessary information to generate a relevant three-month training program, without overwhelming it with unnecessary details.

The key is to strike a balance - provide just enough context to guide the AI, but not so much that it becomes confusing or unproductive. By asking yourself these three questions, you can craft the perfect amount of context to include in your prompt.

The exemplars component is a fancy way of saying "examples". Research on large language models (LLMs) has shown that including examples within the prompt can drastically improve the quality of the output.

Here are some examples of how to use exemplars:

1. Simple Example: If you have a bullet point from a resume, you can ask ChatGPT to rewrite it using a specific structure, such as "I accomplished [X] by [measure Y] that resulted in [Z]".

2. Interview Prep: For an interview question like "What's your biggest weakness?", you can provide the STAR (Situation, Task, Action, Result) framework as an example structure for ChatGPT to follow.

3. Job Description: If you need to write a job description, you can provide an existing job description as an example and ask ChatGPT to reference the formatting and language.

The key takeaway is that exemplars are not necessary for every prompt, but including a relevant example or framework can greatly improve the quality of the output.

The Persona component is essentially who you want ChatGPT and Bard to be. The pro tip here is to think of someone you wish you had instant access to with the task you're facing.

For example, if you're working on a fitness-related task, the Persona could be a physical therapist with experience helping athletes recover. If you're a job seeker, the Persona could be a recruiter or hiring manager. For a creative brief, the Persona could be a senior product marketing manager who's great at storytelling.

You can also name specific individuals, but the results tend to be good only when they're famous enough, like Warren Buffett, Steve Jobs, or Jeff Bezos.

Fictional characters can work as well. For instance, you could ask ChatGPT to draft an email from Batman, and it might even include details like "Please let Alfred know" and sign off as "The Dark Knight."

The key is to visualize the type of expert or authority figure you'd like to have assist you with the task at hand, and then incorporate that Persona into your prompt.

The fifth component of a good prompt is the format. The pro tip here is to literally close your eyes and visualize how exactly you want the end result to look.

For example, if you want a million likes on your thirst trap Instagram photo, you can specify the format as a social media post. If you're working on user feedback, you might want the output in a table format with headers like "Original Feedback", "Team Responsible", and "Priority".

Other common formats include emails, bullet points, and code blocks. However, the one I've found to be the most useful as a full-time working professional is paragraphs and Markdown. For instance, if I receive a lengthy industry report for my director, I can ask ChatGPT to give me the three key takeaways and summarize the report by topic, using H2 as section headers.

Pro tip: Whenever I use ChatGPT to proofread any document, I specify that all changes need to be bolded so I can easily see exactly what has been changed.

The tone component of a prompt is about conveying the desired mood or attitude in the output. It's important to specify the tone you want, as language models like ChatGPT and Google Bard can generate content in a wide range of tones.

Some common tone options include:

• Casual: Use informal, conversational language.
• Formal: Maintain a professional, polished tone.
• Enthusiastic: Express excitement and positivity.
• Objective: Provide a neutral, factual perspective.
• Empathetic: Show understanding and compassion.
• Concise: Use clear, succinct language.
• Witty: Incorporate humor and playfulness.

When determining the appropriate tone, consider the context of the task and the intended audience. For example, a tone that is suitable for a marketing email may differ from the tone used in a technical report.

A pro tip is to provide specific tone keywords in your prompt, such as "use a friendly and approachable tone" or "write in a formal, professional manner." This helps ensure the output matches your desired tone.

Remember, the tone component is the final piece of the prompt puzzle, but it can have a significant impact on the overall quality and usefulness of the generated content.

Dear Tim,

TL;DR We have just unveiled the latest Apple product in collaboration with Tesla - the Apple Car. We have received 12,000 pre-orders, which is 200 higher than our target.

Project Background The Apple Car project was initiated to leverage our expertise in software and design, combined with Tesla's leadership in electric vehicles. The goal was to create a seamless, innovative, and sustainable transportation solution that would redefine the automotive industry.

Business Results The launch of the Apple Car has been a resounding success. We have received 12,000 pre-orders, which is 200 higher than our initial target. This demonstrates the strong demand for our product and the trust our customers have in the Apple brand.

Closing I would like to express my sincere gratitude to the product and engineering teams for their tireless efforts in bringing this project to life. Their dedication and innovation have been instrumental in delivering a truly remarkable product.

Please let me know if you have any further questions or if there is anything else I can assist with.

Best regards, [Your Name] Senior Product Marketing Manager Apple Inc.

In this video, we have covered the six essential components of a well-crafted prompt: task, context, exemplars, persona, format, and tone. By understanding the order of importance and how to effectively utilize each element, you can consistently generate high-quality outputs from language models like ChatGPT and Google Bard.

The key takeaways are:

1. Task: Always start with a clear action verb and articulate the desired end goal.
2. Context: Provide just enough relevant information to constrain the possibilities and guide the model.
3. Exemplars: Include relevant examples or frameworks to help the model understand the desired structure and style.
4. Persona: Specify the type of expert or authority figure you want the model to emulate.
5. Format: Visualize the desired output format and provide instructions accordingly.
6. Tone: Communicate the desired tone of voice, using specific adjectives and adverbs.

By following this comprehensive prompting formula, you can unlock the full potential of language models and streamline your productivity across various tasks. Remember, practice makes perfect, so keep experimenting and refining your prompts to achieve consistently excellent results.