Lesson 8: Guide on best prompting practices

Prompt Engineering framework by 100mentors

Enrich your prompts with specific and relevant information that helps GenAI focus and provide useful responses.
Think about and include the following 6 + 1 elements:

1. Role: Consider who is the most suitable person to answer the question you’re asking. The role can define the level of expertise in the response you’ll receive.
2. Context: What is the general context of your task or project? Providing context helps supply the right information to ensure the answer is relevant.
3. Mission: What exactly do you want to create or accomplish with your prompt?
4. Audience: Who are you addressing? Who will benefit from what you’re trying to create? Defining the audience ensures the language and level of the prompt are appropriate for those responding.
5. Goal: What is the final outcome you want to achieve? Set clear goals to ensure the response is properly directed.
6. Output format: What do you want the final result to look like? What is the ideal form it should take?
7. Examples: Add examples from similar tasks you’ve completed in the past.

To apply the 100mentors prompt engineering framework, follow the structure below:

1. Role: Act as [the most suitable role for the task you want to complete] who specializes in [the main responsibility related to the task].
2. Context: [General information about the organization within which your task is being carried out, or your personal context.]
3. Mission: Develop [a specific deliverable, e.g., a report, a summary, etc.] that will [purpose of the task, e.g., solve a problem, achieve a goal]. Make sure to address [key points that must be included].
4. Audience: This task focuses on [target audience who will benefit from what you’re creating, e.g., personal use, clients with specific characteristics].
5. Goal: [The ultimate aim or impact of the task, e.g., skills development, earning a degree].
6. Output Format: [Ideal format for your deliverable, e.g., a detailed project plan, a table with sections, a one-page summary with word limit, a presentation].

Common mistakes and best practices

Vague Prompts 

Instructions that are not specific and clear do not help GenAI understand exactly what we are asking for.

We need to describe our requests accurately in order to effectively guide the model.

Overly Complex Prompts

Prompts that are complicated and overloaded with too many details make it hard for GenAI to distinguish key information.

It’s better to provide information gradually, in clear and structured thematic sections.

Prompts Without Context

GenAI does not know the context of the task in advance.

We should provide enough information about the purpose of the task and the approach we intend to take to achieve it.

The Idea That „I Need to Know Everything”

It’s not necessary to know all the details in order to pose a question to GenAI.

On the contrary, we can ask for guidance and expand our knowledge through interaction with it, helping us understand what pieces of information we may be missing.

Stopping the Interaction after the first answers

The first responses are not always the best.

Ongoing interaction with GenAI—providing additional input and ideas, as well as receiving its guidance in return—can significantly improve the results.