1. Role Assignment and Context Establishment
Begin by assigning a clear role to the AI and providing detailed context. This helps the AI understand the task and your expectations.
Example Prompt:
“As an expert data analyst and Excel specialist, you will follow my instructions step by step to create and analyze a summary table. Our goal is to summarize complex datasets, extract insights, and provide meaningful visualizations. At every step, you must retain the structure of the existing summary table and the details of previous analyses. Each new instruction will involve adding or expanding the current table, never removing or altering existing columns.”
2. Defining Variables for Columns
Define variables that represent the table columns to help the AI track and manage them efficiently.
Example Prompt:
“We will use variables to represent the columns in our summary table:
Column1 = ‘Product Category’, Column2 = ‘Total Sales’. As new columns are added, these variables will be updated. This will help you manage the table structure clearly.”
3. Initial Analysis Request
When making the first request, use the variables and emphasize preserving prior steps.
Example Prompt:
“Using the dataset below, create a summary table. Use Column1 (‘Product Category’) and Column2 (‘Total Sales’) to display total sales by product category. Retain the current table structure and avoid deleting any columns.”
4. Multi-Layered Analysis Requests
In subsequent requests, introduce new layers of analysis by referencing previous steps and variables.
Example Prompt:
“Add a new column ‘Region’ to the summary table and name it Column3. Then, analyze total sales by region and product category using Column1, Column2, and Column3. Display total sales for each product category in each region. Ensure the table structure is preserved without removing any columns.”
5. Integration of Advanced Techniques
a. Negative Prompting:
Explicitly state what the AI should not do.
“Do not delete columns, alter data, or remove analyses from prior steps.”
b. Example Data:
Whenever possible, provide example data or mockup outputs. This clarifies your expectations.
c. System Messages:
Use persistent system messages to guide the AI’s behavior:
“Always refer to the most updated structure of the summary table.”
“Retain all previous analyses and columns when adding new layers.” umarım işinize yarar