In this post, we will provide you with some tips for creating suitable prompts to define your app, and we’ll guide you with a practical example.
What is a prompt? How to use it to define my app?
A prompt is the set of instructions you give to ChatGPT to perform a task. To create a good prompt, you should consider the following aspects:
1. Define the purpose and objective of the prompt: Before you start, it is crucial to understand what you want to achieve with the prompt. It could be obtaining accurate information, generating creative texts, or performing a specific task.
2. Research and understand the audience: The prompt should be tailored to the audience your app is targeting. You need to know their needs, preferences, problems, and expectations and use an appropriate tone and vocabulary.
3. Choose the right voice and personality: The prompt should reflect the identity of your app and how you want to communicate with your users. You can choose between a formal or informal voice, serious or fun, professional or friendly, etc.
4. Provide context and relevant information: The prompt should give ChatGPT the necessary information to generate a coherent and useful response. You can include data about your app, your industry, your competition, your target audience, etc.
5. Be clear and specific in formulating the prompt: The prompt should tell ChatGPT what type of response you expect and how it should structure it. You can use short sentences, bullet points, tables, examples, references, etc.
Example of defining an app for an eyewear m-commerce
To illustrate these tips, let’s look at an example of defining an eyewear e-commerce app with ChatGPT. We’ll do it step by step, based on user stories. A user story is a brief and simple description of a functionality that a user wants, from their perspective. For example:
- As a user, I want to browse the catalog of available glasses to choose the ones I like.
- As a user, I want to try on glasses virtually to see how they look on me
- As a user, I want to buy glasses online to have them delivered to my home.
If you notice, the user story usually follows the same pattern “As [role], I want [action], for [result].”
For each user story, let’s create a prompt that asks ChatGPT to generate a description of the functionality and an example of how it would look in the app. The prompt will have the following structure:
Role: Indicate the user’s role who wants the functionality.
Context: Specify the app’s name, the type of app, and the goal of the functionality.
Instruction: Tell ChatGPT what type of response you want and how to structure it.
Example or reference: Optionally, you can give ChatGPT an example or a reference of what you want it to generate.
Let’s see how the prompt for the first user story would look:
Role: User
Context: The app is called GlassesOnline, and it’s an eyewear e-commerce app. The functionality allows the user to browse the catalog of available glasses, filter by categories, brands, prices, etc., and view the details of each product.
Instruction: Generate a description of the functionality and an example of how it would look in the app. Use bullet points for the description and HTML code for the example.
Example or reference: You can draw inspiration from this website: [GlassesWorld].
And here’s a possible result generated by ChatGPT:
Description:
The app displays a screen with the catalog of available glasses, sorted by relevance, novelty, or popularity.
The user can filter the catalog by categories (sunglasses, prescription glasses, reading glasses, etc.), brands (Ray-Ban, Oakley, Prada, etc.), prices (from €10 to €500), colors, shapes, materials, etc.
The user can view the details of each product, such as the name, description, price, ratings, photos, etc.
The user can add the product to the cart or continue browsing the catalog.
What do you think? Are you going to put it into practice?
