Prompt engineering is the art of crafting questions or statements that direct an AI model to provide the desired response. Prompt engineering can be used to create highly accurate, relevant, and creative responses from AI models. In the world of generative AI and large language models, prompts are what we use to get AI-generated outputs.

Davis is an experienced instructor who has taught over a million people technology and business skills. He introduces himself and discusses the power of AI and how it can be used for tasks like creating personalized interview prep sheets, writing code, and strategizing. He explains that AI is like a hyper-intelligent business partner that can accomplish a wide range of tasks with the right prompt engineering.

Here, we introduce the concept of prompt engineering and how it can be used with generative AI and large language models to create more accurate, relevant, and creative responses. You’ll find a simple example of how a user might ask a friend for lunch advice and how an AI system could provide a more tailored and informative response based on the user's dietary restrictions and preferences. We’ll also discuss the importance of context and instructions when creating prompts for AI models.

The ICO prompt design framework consists of three main parts: Instructions, Context, and Output. Instructions are used to calibrate the AI model’s interpretation of the prompt. Context provides background information and data that the AI model needs to generate a response. Output defines the desired format of the response.

In this uniquely-formatted interview series, you’ll meet a number of generative AI users from a variety of backgrounds who each share information about how they use in-context learning in their prompt engineering activities.

You are now ready to participate in an interactive learning simulation that will enable you to practice building an ICO-formatted prompt that’s based on a real-world scenario. Head to eazl.link/psim and run through the learning simulation. You can even get an instant certificate of completion for the simulation after you finish!

In this simulation, you will be working on a project for your city government. You are on a committee that is tasked with selecting and integrating a new AI solution that will be used by the city. Specifically, your job on the committee is to represent city workers and consider how they will be impacted by the new AI system.

As you know, AI is excellent at helping to summarize large amounts of data. You have been spending hours reading proposals, bids, and other documents related to the new AI system, and it is taking up a lot of your time. You decide to turn to AI and prompt engineering to help you summarize this information.

In this simulation, we will not address any advanced prompting syntax. We will just focus on the structure of the prompt. So, let’s get started!

In this section, you learned the fundamental concepts of prompt engineering. A prompt is a powerful tool that allows you to harness the capabilities of AI. It is like a key that unlocks the door to AI’s potential. Prompt engineering is the art of crafting effective prompts. You can use prompts to instruct AI models, provide context, and define the desired output. With prompt engineering, you can achieve greater accuracy, relevance, and creativity in AI responses. This can lead to significant improvements in productivity, efficiency, and innovation across a wide range of business applications.

This lecture introduces four tools that can be used to write effective prompts for use with ChatGPT / generative AI work. These tools are delimiters, definitions, Markdown, and handlebars. Delimiters are used to specify the boundary around something in a prompt. Definitions are words or phrases that refer to something that has been added to the prompt. Markdown is a simple way to format plain text. Handlebars are used to create a space for the AI to fill in each time it generates text.

In this lecture, we delve into the powerful techniques of prompt engineering, focusing on the use of delimiters and definitions. Delimiters, such as tags, are sequences of characters that define boundaries around specific information in your prompt. We learn how to use tags to separate or chunk up information for an AI, maintaining a consistent tag structure for clarity.

Definitions, on the other hand, are words or phrases used to refer to something added to your prompt. They simplify the process of referring to large chunks of information in your prompt. By combining delimiters and definitions, we can efficiently guide AI to analyze and understand complex information, such as a lengthy policy document.

This lecture is a crucial part of mastering AI and productivity, as it equips learners with advanced AI prompting strategies. It's a stepping stone towards becoming a proficient prompt engineer, capable of leveraging AI for automation and business productivity.

In this lecture, we will discuss how to use Markdown to structure prompts for large language models.

Large language models are a form of AI that converts the text in your prompt into 1s and 0s. Then, the AI model looks for patterns in its training data to make a prediction, which is your output.

When your prompt is pre-processed and broken into little bits, it’s all plain text. It’s not formatted, so how do you indicate titles and things like that? Through Markdown.

Markdown is a simple, readable way of formatting plain text. There are a number of ways to use Markdown, including creating headers and subheaders, bold text, italic text, lists, and hyperlinks.

In prompt engineering, you’ll most often need Markdown for headers to show the AI which text is a title, which is body content, what’s a sub-section, and so forth.

We use the hashtag symbol for headers, and the approach is simple. One hash is header 1 (the biggest header), two hashtags are the second-biggest header, three hashtags are the third-biggest header, and so on.

You can use Markdown headers to structure your prompt’s contextual information and to guide output formatting. For example, let’s say you want the AI to generate a weekly report for you using some data that’s updated weekly. You’ll add a specific output structure to this prompt by adding Markdown headers to the output part of your prompt.

So, because you’ve added Markdown headers to the output part of your prompt, you’ll guide the AI to generate a weekly report draft that’s formatted how you’d like it to be.

Using Markdown headers can help you write prompts that are more effective and easier for AI models to understand. This can lead to better results, such as more accurate, relevant, and creative output.

In this side-by-side, let's work together to take a document that's structured like many documents that you might encounter in your work and structure it for use with generative AI. In this case, we're going to be taking a publicly accessible education document, a document on the Texas State Expository Essay grading standards, and we're going to convert that into something that we can integrate into a business grade prompt. To do that, we're going to use the skills that you've just learned, delimiters, definitions, and markdown.

In this lecture, Davis introduces the concept of handlebars, which are a powerful tool used in AI prompting. He explains that handlebars are like a stage, where the AI fills in the blanks each time it generates. Davis provides several examples to illustrate how handlebars can be used, and he emphasizes that they are most commonly used in the output section of prompts to guide the AI's generation of dynamic output. He also discusses the benefits of using handlebars, including the ability to have the AI dynamically add words, phrases, or numbers to the output, as well as the ability to give the AI detailed instructions that include markdown, definitions, and nested handlebars.

In this lecture, we’ll review the four essential prompt engineering skills: delimiters, definitions, markdown, and handlebars. These skills will allow you to build prompts that do almost anything.

Delimiters are used to put information between tags and define it with a word or phrase. This enables you to refer to that information easily in your prompt.

Markdown’s hashtag-based headers give structure to your prompt’s contextual information and style your output.

Handlebars are used to create dynamic output based on instructions you add between curly braces ({{ …instructions }}).

These four tools compliment each other and allow you to construct high-quality AI prompts that can be used in your work.

This section introduces advanced prompt engineering techniques. Two easy techniques are introduced to improve AI results, followed by discussions on AI stepping back, multi-agent prompting, and the TESSA SPP prompting approach.

This lecture introduces techniques for crafting effective prompts that leverage AI's emotional understanding and access to specific training data.
Emotional Appeals: AI models respond to emotional appeals, improving results by up to 8%. Incorporate emotional language in prompts, such as "This job opportunity fulfills a long-held dream."
"According to..." Technique: Direct AI to use specific data sources in its responses. Add phrases like "according to data from Wikipedia" or "use information from peer-reviewed research." Benefits: Enhanced AI reasoning abilities, more precise and relevant results, improved productivity and efficiency in business tasks

For further information on emotional stimuli in prompts, visit eazl.link/emotion. For more on the "according to..." technique, visit eazl.link/accordingto. Enhance your prompt engineering skills today and unlock the full potential of AI for your business.

Discover the "stepping back" technique, a proven method to improve AI performance by up to 27%.

Key Concepts: AI models benefit from high-level thinking and iterative analysis. Using "stepping back" prompts, AI can access insights from vast training datasets.

Application Examples: Career planning: Ask AI to consider trends in medicine before generating a personalized plan. Information management: Request that AI reflect on solving information overload before developing a tech strategy.

Benefits: Improved accuracy and specificity of AI responses. Effective utilization of AI's extensive training data. Enhanced productivity and efficiency in business applications.

Harness the proven prompt engineering technique, Solo Performance Prompting (SPP), to supercharge your prompt engineering skills.

Benefits of SPP: Simulate multiple expert viewpoints and perspectives, enhance problem-solving and creativity, quantifiable results up to 20% better.

Using the TESSA Framework for SPP:

• Task: Define the problem to be solved

• Experts: Name the hypothetical personas involved

• Start the Discussion: Initiate the collaboration between personas

• Synthesize: Combine the insights from each persona

• Agreement: Have the personas reach a consensus on a solution

In this section, you learned some advanced prompt engineering techniques. Let’s review them.

• When you include emotional appeals in your generative AI prompts, like “this matters a lot to me”, AI will generate statistically better results.

• Use the “according to…” technique to have AI models use specific parts of its training dataset when generating results.

• If you’re doing AI work that involves specificity, tell the AI to “step back” and recall what it knows about a concept in your prompt.

• For complex problems or synthesizing viewpoints, use the SPP method with TESSA: Task, Experts, Start the Discussion, Synthesize, Agreement.

This lecture explores how prompt engineering can revolutionize the efficiency of industrial mining operations. By crafting precise prompts for AI assistants, businesses can automate the tedious task of reviewing warranty claims for mining equipment.

The lecture begins by highlighting the potential of AI to automate workflows and enhance productivity. It then introduces the ICO framework for structuring high-level prompts, covering elements such as Instructions, Context, and Output.

Next, the lecture demonstrates advanced prompt engineering techniques for mining-specific scenarios. Students will learn how to handle emotional appeals, delimiter tags, markdown formatting, and output handlebars. They will also participate in an interactive simulation to refine their prompt design abilities.

The lecture concludes with practical examples of real-world business prompts, including pre-screening warranty claims and summarizing research reports. Students will also gain access to a library of super prompts for various business tasks.

Start by identifying the goal of the prompt - what do you want it to accomplish? Here, the goal is to categorize warranty requests and provide justifications.

Determine what elements you need to achieve that goal. In this case, warranty request categories and examples are needed for the context. Product details are also required so the AI can properly assess requests.

Structure the output so it is consistent each time - here, putting requests into one of four categories and providing a justification paragraph.

Add instructions last, once you know what the prompt needs to do. Set appropriate AI roles and boundaries.

Leave space to input actual warranty requests to be assessed.

Build natural language connections between the prompt elements (e.g. have the AI "step back" and consider the information).

The process begins with the end goal, then works backwards to identify the required contextual information, output structure, and instructions to achieve that goal. The prompt is designed systematically element by element.

In this lecture, we explore the concept of "context" in prompt engineering for business. We demonstrate how to provide additional information to AI models through prompts, enabling them to perform tasks effectively even when they lack specific knowledge.

Specifically, we focus on the domain of warranty assessment. By leveraging "in-context learning," we enhance the model's understanding of relevant parameters, categories, and catalog information. This approach empowers the AI to assess warranty requests more accurately and efficiently.

Through practical examples and exercises, we guide learners in structuring their prompts using the ICO framework (Instructions, Context, Output), incorporating emotional appeals, delimiter tags, and markdown formatting to improve prompt efficacy.

In this lecture, you will learn how to control the output of your prompts using markdown and handlebars. This will allow you to generate output in a specific format, making it easier to integrate your AI work into advanced systems and workflows. You will also learn how to use emotional appeals and delimiter tags to improve the quality of your prompts. By the end of this lecture, you will be able to use advanced techniques to create high-quality prompts that will help your business leverage the upside of AI while mitigating the downside risks.

In this lecture, you'll learn how to use generative AI models to help you build effective prompts that streamline your business operations and enhance productivity. Let’s explore the use of AI prompts like the one at eazl.link/metainstructions to generate comprehensive instructions for your prompts. By incorporating artificial intelligence into your prompt engineering process, you can elevate your efficiency and unlock the true potential of generative AI.

In this lecture, we’ll use test submissions to see how our warranty claims prompt performs in the real world. You can find test submissions to use yourself in the study guide.

In this concluding lecture, AI expert Davis Jones encourages learners to request their Prompt Engineering Certificate. He emphasizes that prompt engineering is an art, highlighting the numerous ways to craft effective prompts. Jones advises students to embrace creativity and utilize AI to enhance their productivity. He concludes by emphasizing the versatility of AI and the importance of using the right prompts to unlock its full potential in business.