AskElephant prompt nodes

Overview

Prompt nodes are the foundation of powerful AskElephant workflows. These built-in tools allow you to execute AI tasks, filter data, set conditional logic, and process information at scale—all without leaving your workflow. Whether you're automating customer insights, sorting through meeting data, or running bulk analyses, prompt nodes give you the flexibility to customize how AI works within your processes.

This guide covers the four core prompt node types and how to use each one to build smarter, more efficient workflows.

Run Prompt: Your Workflow Foundation

The Run Prompt node is the most versatile tool in your AskElephant toolkit. It executes a single AI task and returns the generated text, making it perfect for one-off analyses without the need to start a full conversation.

What It Does

Feed the node a prompt, select an AI model, and get back immediate results. You can then use those results in other parts of your workflow—making it the connector between your different workflow steps.

Key Configurations

Output

The node returns one variable: AI Completion—the text generated by the AI model. You can reference this output in any subsequent workflow step.

When to Use It

• Summarizing meeting notes
• Extracting key action items from conversations
• Generating customer-facing responses
• Analyzing data in your workflow

Conditional Prompt: Add Logic to Your Workflows

The Conditional Prompt node makes your workflows intelligent by allowing them to proceed or stop based on AI-evaluated criteria. Instead of running the same workflow for every scenario, you can set conditions that determine whether the workflow continues.

What It Does

You define conditions, and the AI evaluates the current workflow state against those conditions. If conditions are met, the workflow proceeds. If not, execution halts—unless you configure it otherwise.

Key Configurations

Example Use Case

You want a workflow to only run on high-priority meetings. In the Conditional Prompt, you'd set conditions like: "only proceed if the meeting transcript mentions 'urgent,' 'critical,' or 'executive.'" The AI analyzes each meeting and moves forward only when conditions are true.

Filter Prompt: Intelligently Sort Your Data

The Filter Prompt node uses AI to sort through lists of objects (meetings, contacts, companies, or calendar events) and pull out only what you need based on natural language criteria.

What It Does

Instead of manually sorting through dozens of records, tell the Filter Prompt what you're looking for in plain English, and the AI handles the filtering automatically.

Key Configurations

Example Scenario

You have 15 meetings from the past month, and you want to find only those that mention "budget planning." Rather than reviewing each manually, the Filter Prompt reads through all 15, evaluates each against your criteria, and returns only the meetings that match.

Output

The node returns Filtered Objects—a list variable containing only the records that matched your criteria. Use this output in the next workflow step or as the final result.

Loop Prompt: Process Data One Item at a Time

The Loop Prompt node runs an AI prompt on each item in a list individually, then collects all the results. This is especially powerful for avoiding AI errors and processing large datasets efficiently.

What It Does

Instead of asking the AI to analyze 10 meetings all at once, the Loop Prompt analyzes them one at a time. This keeps the context focused and reduces hallucinations (AI-generated inaccuracies).

Key Configurations

Why Process One at a Time?

When the AI evaluates items individually, it doesn't get confused by trying to hold too much context at once. This is especially valuable when you're pulling the same type of information from many records—summaries, action items, sentiment analysis, and so on.

Output

The node returns AI Completions—a list variable containing the AI's response for each item processed.

When to Use It

• Generating summaries for 20+ meetings
• Extracting action items from every contact's interaction history
• Analyzing sentiment across multiple conversations
• Processing bulk data without accuracy loss

Putting It All Together: A Real-World Example

Here's how these nodes work together in a single workflow:

1. Run Prompt identifies key topics from a recent meeting
2. Conditional Prompt checks if the meeting involves a specific customer segment
3. Filter Prompt pulls related past meetings with that customer
4. Loop Prompt extracts action items from each filtered meeting
5. Final output becomes a comprehensive customer brief

Each node passes its output to the next, building a powerful multi-step analysis without manual work.

Next Steps

Now that you understand the core prompt nodes, explore how to combine them in your own workflows. Start simple—use a Run Prompt to test your first task, then layer in conditional logic, filtering, or looping as your needs grow.

Have questions about which node fits your use case? Connect with the Herd in our community forum to see how others are using prompt nodes in their workflows.