Prompt Best Practices
This guide explains how to effectively interact with Flatlogic’s AI Engineer to get the best possible outcomes from your prompts. Clearly structured prompts help you precisely express your intentions, save time, reduce errors, and maximize the efficiency of your application development.
Who Is This Guide For?
Product owners, project managers, and business analysts who specify requirements.
Developers looking to efficiently communicate app changes or new features.
Any users interacting directly with the AI Engineer to manage or build their applications.
Why Prompt Clarity Matters
The AI Engineer relies on clear and explicit prompts to accurately interpret your goals. The more precise your prompts, the fewer iterations required to achieve the desired result.
Prompt Best Practices
Follow these guidelines to optimize your interaction with the AI Engineer:
1. Be Clear and Precise
Clearly state the functionality or change you desire.
Ambiguous prompts lead to misunderstanding and suboptimal outcomes.
Good Prompt:
“Create a 'Product' entity with 'Name', 'Price', and 'Description' fields.”
Poor Prompt:
“Set up products.”
2. Describe Entities and Relationships Clearly
- Explicitly mention the structure, fields, and relationships between entities to achieve accurate results.
Good Prompt:
“Create an 'Order' entity linked to 'Customer' with fields: order_date, amount, and status.”
Poor Prompt:
“Make orders with customers.”
3. Provide Concrete Examples
- Clarify your intentions using real-world examples for fields and entities.
Good Prompt:
“Include a 'Payment Method' field with options like Credit Card, PayPal, and Wire Transfer.”
Poor Prompt:
“Add payment info.”
4. Split Complex Requests
- Break down complicated requirements into smaller, logical steps, to simplify understanding.
Good Prompt:
“First, create a 'Customer' entity, then add fields for 'Address' and 'Phone Number'.”
Poor Prompt:
“Build a CRM with customers, leads, orders, invoices all at once.”
5. Request Incremental and Iterative Changes
- Incrementally enhance your schema or app, verifying each step before proceeding further.
Good Prompt:
“Add a 'Discount' field to the existing 'Invoice' entity.”
Poor Prompt:
“Ensure invoices include all standard fields.”
6. Verify Results Frequently
- Regularly confirm and validate outputs, ensuring alignment with your expectations before moving forward.
Good Prompt:
“Show me the schema for review before generating screens.”
Poor Prompt:
“Generate everything first; I'll check later.”
7. Avoid Jargon or Vague Terms
- Use straightforward language; clearly define industry-specific or ambiguous terms.
Good Prompt:
“Create a field called 'Order Total' calculated as Quantity multiplied by Unit Price.”
Poor Prompt:
“Set order amount correctly.”
Example Workflow
A typical interaction might look like this:
Step 1: Clearly state your initial need
“Create an 'Employee' entity with fields: First Name, Last Name, Email, and Hire Date.”
Step 2: Verify and refine
“Show the created schema and confirm entity relationships.”
Step 3: Proceed iteratively
“Now generate CRUD screens for 'Employee' entity.”
This workflow ensures precision, minimizing the need for corrections or revisions.
Summary
Adhering to these prompt best practices ensures effective use of Flatlogic's AI Engineer, streamlining your app-building process and ensuring clear, accurate, and reliable results.
If you face any difficulties, please message us on our Discord, Forum, Twitter, or Facebook. We will respond to your inquiry as quickly as possible!