In Narrativa® Navigator, prompts are not just instructions for the AI. They are the connection between your data and the content you want to generate.

Knowing how to write strong prompts helps you:

• Get more accurate results
• Reduce manual editing
• Make better use of your datasets
• Automate content with full editorial control

What is a prompt?

A prompt is the instruction you give to the AI to generate a response. It can be a question, a command, or a description of the desired output.

The quality of the prompt directly affects the quality of the result.
A clear prompt leads to better content.

In Narrativa® Navigator, prompts can also connect directly to your data using dynamic references like #field. This allows you to generate personalized content at scale.

Prompts + data: the core of Narrativa® Navigator

Unlike general AI tools, prompts in Narrativa® Navigator are data-aware. You don’t write them in isolation—you write them to interact with structured data.

Here’s how it works:

• Each dataset contains fields (columns).
• You can reference these fields in the prompt using #fieldname.
• The AI uses those values to generate tailored content for each row.

Example:
If your dataset includes a column called temperature_max, your prompt might say:
“The maximum temperature forecast for today is #temperature_max degrees.”

The text changes automatically based on the data in each row.

The prompt defines the narrative; the data shapes the content.

Writing prompts in Narrativa® Navigator vs. ChatGPT

Narrativa® Navigator is not a chatbot. It’s a professional, data-driven content automation platform. This changes how prompts are written—and what they’re for.

In ChatGPT:

• You write a prompt
• You get a single response
• You copy, paste, and repeat manually

In Narrativa® Navigator:

• You write a prompt once
• It runs automatically across hundreds or thousands of records
• Each result uses different data
• The output is structured and ready to publish

In Narrativa® Navigator, the prompt is not a question. It’s a smart content template.

Key principles for writing effective prompts

These best practices apply everywhere—but are essential in automation:

• Clarity: Be specific and avoid vague language.
• Context: Explain the goal or use case of the content.
• Instruction: Use clear verbs like write, summarize, generate, compare.
• Format: Define how the output should look—list, paragraph, headline, etc.
• Tone and style: Specify the desired voice—journalistic, technical, educational.
• Roleplay: Ask the AI to act as a persona, e.g., “Act as a data journalist…”
• Iteration: Test and improve until the result meets your needs.

Where to create prompts in Narrativa® Navigator

You can write prompts in three main areas of the platform:

• Data: Inside AI Agent Scraping features.
• AI Agents: For Automatic Writing, Advanced Translation, Web Scraping, and Semantic Search.

Ideal prompt structure in Narrativa® Navigator

Use this simple and effective structure:

• Context: What is the purpose of the content?
• Data: Which dataset fields should it use? (e.g., #city, #sales_total)
• Task: What should the AI generate?
• Format: How should the output be presented?
• Tone / Style: What tone should it follow?
• Restrictions: What should the AI avoid (e.g., length, jargon, repetitions)?

Practical template

Context: [Where or why the text will be published]  
Data: Use #field1, #field2, #field3  
Task: [What the AI should generate]  
Format: [Paragraph, list, summary, etc.]  
Tone/Style: [Informative, journalistic, technical]  
Restrictions: [Avoid repeated values, limit to X words]

Tips for prompt success in automation

• Use data only when it adds real value
• Design for scalability, not one-off tasks
• Avoid ambiguity—be precise
• Break down complex tasks into smaller parts

Why good prompts matter

Writing better prompts means:

• Higher quality content
• Fewer revisions
• More control over output
• Scalable personalization
• A stronger connection between your data and your message

In Narrativa Navigator, the prompt is your invisible editor.

Your data says what’s happening; the prompt decides how to tell the story.