Skip to main content

​
Basics of prompting

Effective prompts are the key to better results with BusinessGPT. A well-structured prompt leads to more accurate, relevant, and useful responses.

​
Structure of a good prompt

​
1. Clear role assignment

Define the role that the AI should take on: 
You are an experienced marketing expert. Analyze the following campaign...

​
2. Provide context

Provide relevant background information: 
Our company is a B2B software provider with 50 employees.
We develop CRM solutions for small and medium-sized businesses.

​
3. Define a specific task

Formulate your request precisely: 
Create a 10-point checklist for introducing our software to new customers.

​
4. Specify the desired format

Specify how the answer should be structured: 
Format the answer as a numbered list with a short explanation for each point.

​
Proven techniques

​
Chain of thought (step-by-step thinking)

Ask the AI to explain its thought process: 
Explain step by step how you arrived at this solution.

​
Few-Shot Learning (Provide examples)

Provide examples of the desired result: 
Here are two examples of good product descriptions:
1. [Example 1]
2. [Example 2]

Now create a similar description for our product.

​
Iterative refinement

Refine your prompts based on the results: 
That's good, but can you make it more specific to our target audience?

​
Avoid common mistakes

​
❌ Too vague requests

Bad: “Write something about marketing.” Better: “Create a 5-point strategy for social media marketing for B2B companies”

​
❌ Missing context

Bad: “How can I improve this?” Better: “How can I improve this email campaign? [Insert email text]”

​
❌ Multiple tasks in one prompt

Bad: “Analyze the market, create a strategy, and write a business plan.” Better: Break this down into separate, focused prompts.

​
Special techniques for BusinessGPT

​
Use knowledge bases

Refer to uploaded documents: 
Based on the uploaded manual, explain the most important features of our software.

​
Optimize model selection

  • GPT-4o: For complex analysis and creative tasks
  • Claude: For nuanced text processing and longer documents
  • Faster models: For simple questions and summaries

​
Use slash commands

Use /[search term] to quickly access prompts and knowledge bases.

​
Prompt templates

​
Analysis template

Role: [expert role]
Context: [background information]
Task: Analyze [element to be analyzed]
Focus: Concentrate particularly on [specific aspects]
Format: Structure the answer as [desired format]

​
Creation template

Create [type of content] for [target audience]
Style: [desired tone/style]
Length: [approximate word count]
Structure: [desired outline]
Special requirements: [specific specifications]

​
Revision template

Revise the following text:
[original text]

Improvements:
- [specific improvement requests]
- [additional requirements]

Target audience: [description of target audience]

​
Advanced strategies

​
Persona-based prompting

Create detailed personas for consistent results: 
You are Sarah, a 35-year-old marketing manager with 10 years of experience
in the tech industry. You are analytical, but also creative, and pay
particular attention to ROI optimization.

​
Constraint-based prompting

Set deliberate limits for more focused results: 
Explain the concept in a maximum of 3 sentences, do not use technical terms,
and address it to people with no prior experience.

​
Multi-step prompting

Break complex tasks down into steps: 
Step 1: Analyze the data
Step 2: Identify the 3 most important trends
Step 3: Develop recommendations for action for each trend

​
Quality control

​
Evaluate results

  • Is the answer complete?
  • Does it match the desired format?
  • Is it appropriate for the target audience?
  • Are all important points covered?

​
Iterative improvement

That's a good start. Can you elaborate on point 3
and add a concrete example?

​
Tips for different use cases

​
Business documents

  • Use professional language
  • Structure clearly with headings
  • Ask for sources for facts

​
Creative content

  • Set the style and tone
  • Use inspiring examples
  • Experiment with different approaches

​
Technical documentation

  • Define the experience level of the target audience
  • Ask for step-by-step instructions
  • Request examples and code blocks

​
Data analysis

  • Provide the context of the data
  • Define specific analysis goals
  • Ask for visualizable results
By applying these best practices, you can significantly improve the quality and relevance of BusinessGPT’s responses and work more efficiently.